Skip to main content
OCLC Support

2019 AskQC office hour member Q&A

Review all AskQC office hour member questions from 2019.

January 2019: When to input a new record

Topic-specific questions

If a separate record, an M record already exists, should we report the M record?

Yes, please do report and we will take care of that. Please note that we do have a significant backlog of reported duplicates. So when you do report a duplicate to us, know that we put it in our queue, we will get to it eventually, but we can't guarantee that we will get to it right away. It might be a number of months before the duplicates are resolved, but we always appreciate reports of duplicate records.

We've noticed over the last few years something that seems to be standard practice, but which we cannot find documented. It seems that LC is treating some publishers' statements of "paperback edition" not as edition statements, but as printing statements, thereby negating the need for a new record. Is this OCLC practice as well?

Yes, and that is documented in When to Input a New Record as well as Differences Between, Changes Within. All other things being equal, that is the pagination, and the size, etc., paperback editions that are stated as paperback editions are generally not considered to be separate editions in this context.

Does OCLC have guidance on advance review copies/uncorrected proofs? Should these be given separate records?

Yes, we do have guidance in the 250 Edition Statements section of BFAS. Those should be separate records. If there isn't a statement that you can transcribe as an edition statement, that is a legitimate edition statement to supply by the cataloger.

In the last example of the presentation for accompanying material, why would a new record be created, rather than updating the existing record? They appear to be the same item, to me, anyway?

The upper record on that slide is the book all by itself without the audio cassette. The bottom record is the book accompanied by the audio cassette. Those are legitimately separate records.

Why does a difference in country of publication require a new record if everything else is the same?

A lot of this has to do with historic publication patterns. It used to be that there was a significant difference between things published in the U.S. and things published elsewhere, such as Great Britain. That difference and distinction is increasingly less in the more global situation in which we find ourselves today, but it used to be a big, big difference. It could be that the cover art was totally different or something about the design of the book was totally different, even if the context of the item was the same. That is something that we decided on historically and we've stuck with it. Maybe it's something we need to reconsider, but for now that is the way we are considering those legitimately separate.

There are some new printings of music textbooks that list URLs for related audio contents. Should these have separate records?

If these are online, you can catalog these separately if that is your policy. You would also want to make note of the presence of online accompanying material on the record for the score itself. 

If you have an item and the only record is in a different language, do you create another record in your language?

Yes, these are called parallel records. BFAS chapter 3, section 3.10 Parallel Records for Language of Cataloging provides guidelines for how to handle these types of records. There was also an Office Hours presentation given in October 2018, for which the recording, questions, and notes are available. 

Following from the discussion of "edition statements" that are really printing statements: If the WorldCat record in WorldCat has what I'm quite sure is a printing statement in the 250 (e.g., "4a. ed."), should it be left as-is? If I have the "1a. ed." and everything else is identical, should I replace the 250 in the WorldCat record with "my" edition statement or leave it as-is?

You could remove that 250 if you are absolutely certain that it is a printing statement and replace the record.

In the past, I cataloged sets of curriculum materials which had extensive supplementary materials, which we usually did not have all of. When libraries appear to be able to pick and choose among supplementary materials, how many separate records for subsets are appropriate?

One thing you are always free to do is to use the record that is in the database and edit it locally to describe what you have in hand or add the local information to an LBD or LHR. If you need a record in WorldCat to describe what you have, and the accompanying material differs from what is in the WorldCat record, then you could create a new record.

Does OCLC have an official document of publishers and their imprints so that records do not duplicate? If so, where can I find that document?

Built into Duplicate Detection and Resolution (DDR) are a relatively limited number of equivalents among publishers. The most obvious example would be what used to be called the Government Printing Office, now called the Government Publication Office, the GPO, the Superintendent of Documents, all of those are different ways that what is now the Government Publication Office have been designated for certain government publications from the United States over the years. So, we have a whole set of equivalents for that particular entity. There are some other publishers for which we have tables of equivalents, as the example of John Wiley versus Wiley and various other variations of Wiley as a publisher. It is not a document and not available publicly, but we do have some of that built into DDR.

Jay mentioned cleaning up print records with 006/007 for 'e'. When these are hybrid records, such as a score that has been digitized and the digital version represented only by the print fields 856, 5xx, 006, 007, are you splitting into two records or leaving these alone?

This goes back to what the history of cataloging electronic resources has been, where you have long had an option to note the existence of an electronic version on the corresponding record for the print. The way to think of these is not so much that they are hybrids of both, but that it really is still just the print record that is making a note of the existence of the electronic version. Of course, that means that print record will have an 856 field, it may have some kind of 5xx field referring to the electronic version, and this problematic 006/007 which is what causes the problem in terms of truly identifying what this record is and the way that they're displayed in systems. When we clean these up, what we'll eventually do is take the 006 and 007 out of the record, but the 856 that notes the existence of the location of the electronic version, that would stay intact. Now in any of these cases you may have this record for the print, a separate record for the electronic version is certainly appropriate and may already be there. But if it's not there, we are not doing anything to remove the 856 fields in these cases. 

I have seen textbook records in which multi-volume sets and loose-leaf volumes are on the same bibliographic record as a one-volume hardbound edition. Shouldn't each of these have their own bibliographic record?

Yes, they probably should have separate records. Something published as a one-volume and the same thing published as multiple volumes and the same thing published as loose leaf would all have separate bibliographic records, justifiably. If there is just a difference in binding, such as spiral bound versus hardbound, but the content is the same then those would not justify separate records. If you have loose leaf versus hardbound, it may be that it's just issued in two formats and is exactly the same situation as described with spiral bound versus hardbound. Loose leaf would have a separate record if, in fact, it was intended to be updated but if it's just loose leaf so that it lies flat on the table, that would not justify a new record.

OCLC BFAS 4.2 says that variation in copyright dates does not justify a new record if the publication dates are the same. Assuming everything else is the same, no publication date, when do we need a bibliographic record for the later date [1963] in the following case: 1. "copyright 1937, renewed 1963"; 2. "c1937, copyright transferred 1963"; 3. "c1937, c1963"?

Some years ago, I had occasion to do some research on copyright, copyright transfers, copyright renewals, all of those copyright variations. The conclusion that I came to in all of that research and what AACR2 at the time and what RDA currently says, is that all of those copyright renewals, transfers, assignments, and so on are essentially to be disregarded. That is documented in AACR2 and fairly certain in RDA or related policy statement. Generally, ignore copyright renewals, copyright transfers, copyright assignment, and all of those other copyright variations. Only what we may call a straight copyright date would be paid attention to.

When a self-publishing author switches companies does that justify a new record? Isn't it just a new printer?

It would probably justify a new record, as is cataloger's judgment. If you have both in hand and can compare them, that would be a way to make the decision. All of the self-publishing and print-on-demand publications fall into this.

Since you mentioned print-on-demand, I have a stack of reprints that we purchased through a print-on-demand publisher. I see some records that use the publisher and date of publication of the original work as well as records that use the print-on-demand publisher and date. Which is correct?

This is really a case where you would want to follow the print-on-demand and photocopy cataloging guidelines, which you can find in the LC PCC Policy Statements for RDA instruction 1.11. So, if you go there it will tell you what to do with these print-on-demand type publications. You would end up retaining as the 260 or 264 the publisher of the original publication and pretty much ignore the publisher of that print-on-demand photocopy that you have in hand.

What is the current policy/recommendation on institution-specific 856 fields in WorldCat records that do not work for users elsewhere because they require local authentication? For example, "Table of contents available to XXX-affiliated users at ... ". Should/can we delete these from the WorldCat record? These cause major problems when they come into our local catalog because they create confusing false links our users can't access, and our reference librarians complain about them.

There are instructions in BFAS under field 856 about what should be done in those cases, and there might even be something mentioned in When to Input a New Record as well. Generally speaking, if those institution-specific URLs can be generalized, often by removing the proxy information, they should be generalized. The institution-specific URLs should not be added to a WorldCat record and you are welcome to try to generalize them according to those instructions in BFAS or delete them. There is no need to keep them in the WorldCat record if they are truly local links. When we encounter them here, we also try to generalize them or delete them.

Jay mentioned the cardinal rule for determining duplicates when cataloging versus when batch loading. Is there a cardinal rule for provider-neutral online serial publications? Are there cases where differences justify a separate record?

It is the case with serials you sometimes tend to be a little more lenient and overlook differences when you're dealing with the same serial that has been digitized independently from issues from different sources. You may have a case where one provider has digitized the articles only, and the other provider has digitized the editorial preface, the advertisements, all of the illustrations, and the whole thing. And that, still within CONSER practice, has ended up on one record with the idea that it could change over time. One of these providers could go back and pick up the content that they were missing otherwise. It's probably something where you have to use some judgment, it may depend on whether the serial is rare or if it's just an ordinary run-of-the-mill serial. When you think about how serials are digitized on Google, where you've got all of these separate individual issues and who knows necessarily what they digitized versus another serial, same title, that has been very carefully digitized, where they did cover to cover and absolutely everything in between. You might think should those be put on separate records, and the approach is generally no, put it all on one record and overlook some of these little differences in where there's a difference in content.

I have some government publications that have been reprinted and they use the original publication date (in the 1950s). These are clearly newer (2000+) but I think I have to use the original record. Is that true? Should I add a note in the record indicating our copy is a reprint?

Generally speaking, if you have a simple reprint you should use the record for the original publication. If there are any differences such as a new bibliography or something new, some other kind of change, then you are encouraged to add a new record with a new date. If it's just a reprint use the original record. You may want to add locally a note to your record or add the local information to an LBD or LHR, saying that the printing is a contemporary print rather than an older print.

If we have a score with a cover title that matches the existing record, but it also has a title page that does NOT match but is very close, and the existing record has no note saying that the title comes from the cover, yet everything else matches exactly, does this require a new record? This seems to happen a lot in music publishing, where new printings appear with covers or title pages that weren't there before, but the music itself and the publication information is identical. I'm never sure what to do. I often find myself wondering whether someone simply didn't add a note about the title source, or is it actually a new publication?

That is one of the great mysteries of music cataloging because there can be so many different sources for titles and other kinds of information - the cover, title, the caption, and so on. You kind of have to use judgment. As time went on, cataloging instructions for cataloging scores were more explicit that you should indicate in the body of the record where a title came from, and of course in RDA you would do that more often. You really have to use judgment, if it's close enough and everything else being equal, I would probably use the existing record. 

Speaking of "print-on-demand", how should one deal with the Smithsonian Folkways CD reprints? It is common to find a record where only the reprint date is different. Can I just use the record and change the date to correspond to when it was reprinted for our order?

CDs reproduced on-demand should probably be treated similarly to other on-demand reproductions; under RDA, that would be following the print-on-demand section of RDA LC-PCC Policy Statement 1.11. If there's an existing record for the standard publication in the same physical format (audio compact disc), you may use that record and either edit it locally for your own copy or document the local specifics in your own Local Data Record. If there's already a record for the on-demand reproduction, use that record for all on-demand reproductions of that manifestation, regardless of the dates of the reproduction. Adapt the instructions in PS 1.11 to the audio format as appropriate.

The Center for Military History is always reprinting government docs published in the 1950x, but they give a new title page. They give a new title page with a new publication date. Would this be a new record? The content is identical, but the title page verso says, "first printed 1951", for example.

My suggestion would be to take the title page date as the date of publication and create a new record if one for that date of publication does not exist.

Sometimes a paper edition comes out after 10 years of the hard copy. Is this considered a major change for a new record?

If the stated date of publication or any other element differs from the original hardcover (including a size difference of more than two centimeters or a pagination difference of more than three pages), a new record is justified.

General questions

I've been noticing records for print books, based on 300 and 3xx fields with 006 and 007 fields indicating they are electronic, and I'm not sure what is happening here. We often have holdings set on them, so I'm assuming they were once clearly print, but now are displaying as 'e'. Any insight as to possible causes/solutions?

We are aware of this problem and are in the process of cleaning those up. Some of that has to do with changes in policy regarding URLs and related records for online versions and print versions. A record for the print version should not have a 006 or 007 field to indicate the presence, or the existence, of an electronic version. It may have a 006 or 007 field for other reasons, for instance accompanying material, but not to indicate the existence of an electronic version.

When I try to control name headings, I often get an error message about a script that can't run, so I can't see the actual NAF record. I end up either selecting insert name and then clicking on it to see the NAF record, or go to LC's authority file. Why does this error happen so frequently?

This sounds like a software issue. We suggest that you send a query, with perhaps a screenshot of the error message you're getting to our Customer Support staff and they may be able to help you.

I'm trying to make more use of the NACO functionality in Record Manager, but I keep running into problems. I often get a system error when I try to add or replace an authority record. Also, the "Copy Authority Data" feature sometimes works, sometimes doesn't. Is the NACO functionality still a work in progress?

NACO functionality is brand new in Record Manager as of the last quarter of 2018, so yes, it is still a work in progress. There are a lot of enhancements that are needed. The basics are there, but there is still work ongoing to make it have the same functionality that you currently have in Connexion. You would want to contact Customer Support about this.

February 2019: Provider neutral cataloging

Topic-specific questions

Is the policy for photocopies and POD resources contrary to RDA?  I would expect to catalog the digital manifestation, and supply notes about the original publication.

Yes, you are correct that is exactly what RDA would have you do. It does run contrary to what RDA says. It's a case of continuing that policy of cataloging a photocopy or Print on Demand publication using a reproduction note where you describe the original and that is completely opposite of what RDA would have you do.

I often see records for electronic resources based on the information on the provider's web page, not the title page of the resource. Should these records be updated to conform with the title page or should a new record be created? Sometimes publication date differs.

It really should be the case that the description is changed to match the title page, because that title page, presumably belonging to more than one instance of that resource, would really be the better choice of the description so that multiple sites could be added to the one Provider-Neutral record.

For print reproduction/POD P-N records, what do you recommend for 300 $c in cases where there is no print original? Can we leave 300 $c out completely?

Certainly, if you don’t know what it is and can’t supply it, it doesn’t need to be included. It is also worth pointing out that in the reproduction note where you might have included the size of the reproduction since it is going to cover all reproductions, that information is not there.  So if you think about how in the past you might have had a thesis that was 28 cm in its original form but you got a photo copy of that same thing from UMI ProQuest in the past and it was shorter it was only 22 cm and you used to see that in records where the 22 cm would have been listed in field 533 as part of the reproduction. That’s not there anymore. You can have really large reproductions or short reproductions; they are covered by the one record. If you know the size of the original include it, but if you don’t know leave it out.

With a digitization, by the owner-library, of a pre-1801 book, how would you code the dates in 008?

I don’t see that as being any different a situation if the book was published 1950, 1850, or 1750. You are going to record the date that is on that original publication that you are finding on that title page even though it makes it look, in theory, that this online version was issued in 1850, which we know is not true. That is just one of the quirks of Provider-Neutral cataloging. 

How do you handle if the only URL you have is a local one?

A local URL is probably better than no URL at all. So, if that is all you have, that is what I would include. If you could figure out something that could be used by others to at least get you to the front door of the resource, I would include that. But if all you have is a local URL, go ahead and include it.

I just cataloged a copy of a handwritten music manuscript that looked like it might be a photocopy, that was sent to binding. Does that qualify for P-N?

If it is just a case that it is a photocopy of a manuscript, you would follow the photocopy and Print on Demand guidelines and describe the original manuscript and put that print reproduction note in a field 533. If something like that were put online, it changes the type code because everything that is online is considered published. So, if this was a manuscript score that would have been Type 'd' originally, once it is online it has to be Type 'c' because it is now a published score.

Can P-N records be coded as PCC?

Yes, because this is a PCC policy, but it is a PCC policy which OCLC has adopted for everything that falls into the online category as well as photocopy and Print on Demand reproduction.

Print on Demand books vary wildly in pagination and illustrative matter included (or not) and named publishers. Are we just supposed to modify a Provider-Neutral record locally? What happens with OCLC updates in this case?

Well, you maybe have some interesting situation there. It is hard to make some statement “Yes, you would always put it on one record if the paging is much different”. You would expect in the normal case you would have Print on Demand reproduction or a photocopy, and if you told someone to go to page 29 and look at the second paragraph you would all be looking at the same thing. But in some cases that is not going to be true because it may be the same content, but somehow laid out differently. In that case I would say it is not really the Print on Demand reproduction it’s just another version of that same resource, probably you need more than one record in that case.

OCLC documentation for When to Input a New Record says that absence or presence of 040 $e does not require a new record. I have been working on streaming videos. Would best practice be to make an existing record PN, even if it requires a lot of major revisions and may have a lot of holdings?

Yes, I don't see where that makes a difference. In the case of, for instance, the photocopy and Print on Demand, since that's a relatively new approach to cataloging that kind of material, we've got 30 years of records going back where we have 533 fields that say "Photocopy. Ann Arbor, Michigan : University Microfilms, etc. and 22 cm." Something like that, you would want to fix it up, make it Provider-Neutral so that it conforms to current standards. It really wouldn’t matter how many holdings are on it.

Could you say more about what differences in provider versions would justify separate descriptions rather than Provider-Neutral treatment, i.e. accessibility differences, content differences, etc.?

Well, you do run into situations with online resources where of course there’s going to be differences in accessibility and that kind of thing. Normally that is just to be over-looked. However, there are more complicated kinds of things where you can’t tell for sure if the title that is available on this one website is really the same thing as available over on this other website. Thinking back, some of the stuff that ended up on that Project Gutenberg website, it would just have a title and it had no indication of any kind of bibliographic history. And because it was plain text you had a hard time being able to tell what had been used as the basis for making this digitized copy. So, if something is digitized and you have got the original title page, then you can say this is all based on the 1859 edition of this work. But over here there is the same title with no information at all as to where it came from, you are going to end up with two records in a case like that, because you cannot tell whether that second item really belongs on the first record.

Other kind of things like file sizes, the file types, that kind of thing get over-looked. PDF, HTML, and plain text could all exist on the same record if it really is the same content.

I am occasionally asked to catalog locally printed copies of electronic resources. Should I make a new record for these printouts with a 533 Print reproduction note?  Wouldn't this imply the publisher printed it even though it was only published in electronic form?

It seems that a locally printed copy of an electronic resource should be Provider-Neutral with a print reproduction note. It is falling into that same category of photocopy or Print on Demand reproduction and still sounds like it would be provider neutral. If you plan to contribute a record to WorldCat, we would want it to be provider neutral. Another option would be to make a copy for local use only, in a local holdings record in your local system, and note that particular copy is a photocopy.

If you're creating a PN record for a digitization of a resource that was unpublished in its original format, do you create a 264 field for publication information?  RDA considers all online resources as published, which would seem to require 264, but PN guidelines say take the publisher information from the original.

Yes. Provider-Neutral works really well for items that were previously published. Especially when you have a title page with the publisher listed on it so that you can transcribe it. But because rules in AACR2 and RDA both indicate that online resources are published, if you have something that is manuscript that is put online it is now published and you should have a publisher. You see this a lot with dissertations that are put online that are cataloged Provider-Neutral where what happens is that you end up with a place and publisher that relate to the library that did the digitization and put the resource online.  I suppose if you are digitizing a manuscript that is 300 years old, and you have some date like 1700, and you are now putting the library in as the publisher of this thing, that is kind of an unusual situation and maybe not so easily answered.

For the unpublished digitized title, wouldn't there be a 264 anyway for the manuscript date? I'm assuming the unpublished item was a manuscript, letter, or diary.

Yes, there would have been a 264 with second indicator of '0' with just the date of the manuscript and possibly the producer of the manuscript. The thing is that once it's online, now it's published. You have the problem that you really ought to come up with a place and publisher for this published online resource.

Does the photocopy/POD apply to self-published works? Our library has donations from a local writer who donates one of her books quarterly.

You might ask is photocopying really the production mechanism to produce copies of the book that is being distributed in which case, I wouldn't pay attention to it. I would take the person's name that is there as the self-publisher at the foot of the title page as the publisher and not put in a print reproduction note, in a case like that. You would only want to do that when something looks like it has a previous existence, because it really is about photocopies of something that already existed or Print on Demand reproductions. So, it really is about reproductions, as opposed to photocopying was the initial mechanism to create copies to distribute.

Would theses and dissertations be cataloged as P-N records?

Yes, they would. It’s online and the Provider-Neutral guidelines for online resources apply to everything. You may think that it won’t be online anywhere else, but with anything electronic, it is hard to say never. If it does end up online somewhere else, the record should already be set up to be able to just add another URL.

Should we add statements and data such as closed captioning in a PN record?

That's one of the issues with trying to apply Provider-Neutral guidelines to videos.  If there is that kind of difference between what is otherwise the same video on two different websites, maybe that does really require two different records to be able to handle that correctly. It's hard to say definitively whether something has not been mentioned in a bibliographic record, such as closed captioning because it is not there, that is it's not in the resource. Or, if it has been overlooked by a particular cataloger. It is hard to know. You have to use your judgment. If you know that there is a version with closed captioning and another without closed captioning, and there is a bibliographic record that corresponds to one of those, and you have access to the other version, then you would create a new record.

It may be worth pointing out that similar kinds of things happen for print publications as well, that end up being digitized after the fact. This came up in serials at an earlier point where if I have this Provider-Neutral record for cataloging a serial, and on this website they actually digitized everything cover to cover but over on this other website they only did the articles. They left out the letters to the editor and advertising, maybe they didn't even produce the photographs that were there. The approach there was to put in on the one record with the idea that this type of thing could change over time. For monograph catalogers, that is kind of a difficult thing to accept, because in terms of print records we would typically make two records if there were those kinds of differences. So, it really is a thing where you have to use your judgment about whether another record is necessary.

Can you repeat the best practice for streaming video again?

The problem, it's not merely streaming video, it has been a long-standing problem with even tangible video recordings such as discs and even before the advent of discs, video cassettes, and so on. The thoroughness of cataloging varies a lot, and sometimes a cataloger will indicate all sorts of information including accessibility information, most commonly as subtitles or captioning or audio description and things like that, and sometimes a cataloger will overlook that or not include that information. The Provider-Neutral guidelines hasn’t really changed how you would determine whether to input a new record or not input a new record. You have to make a judgment about how likely it is that the information that is in an existing bibliographic record is accurate. You also have to consider whether the publisher was accurate in indicating things like captioning and subtitling and all that as well. I’ve presented entire 8-hour workshops on this, and it is hard to sum it up in a few minutes. The problem hasn’t changed. It’s the same problems we have always had that as in it has been exacerbated by the advent of streaming media.

If we add data about our digitized copy in a 506, 533, and 538 with our institution code in subfield $5 in our local bibliographic records (Voyager) and later send those records to OCLC, will those fields be stripped out of the OCLC records? Should we add these fields only after our records have been sent to OCLC?

If the record was going to be added as a new record to WorldCat, all of those fields would remain intact and would be added. If it found a match in the database, we transfer those under certain circumstances. For some of those fields, if they have a subfield $5 that does not match a subfield $5 for the same field that is already in a record, yes, it will transfer. You do not have to add these fields only after the record has been sent, but that is an option.

What about "pages" versus "leaves" or "volume" versus "sheet" for print reproductions for items that don't have a print original? Is there anything special or things to note about this?

If there gets to be a difference in paging versus leaves as a result of the thing being photocopied or going through a Print on Demand process, that gets overlooked. As long as you are ending up with what is essentially the same resource. This kind of thing would happen in the past with the dissertations you used to get from UMI, where the photocopy was pages but the thing that had been photocopied originally was presented on leaves.

If one is cataloging a physical item that might or might not be a reproduction, would it be acceptable to catalog it following P-N guidelines and instructions, but omit Form: r and "Print reproduction" in the 533 field?

No, because if you were not able to conclude that what you had was a reproduction so that you didn't want to code Form: r, then the Provider-Neutral guidelines would not apply to that item. You would catalog the resource in hand as is.

I'm finding records for resources that are published by university presses that have an 's' in the GPub character place. Bibliographic Formats and Standards indicates that publications from university presses are not considered government publications, so this character position should be 'blank'. Some examples are 32969382 (University Press of Kentucky) and 756594353 (University of North Carolina Press). These are DLC records. Does LC have a different practice of coding university press publications as government documents?

No, they no longer have a different practice. There was a change in practice regarding the coding of government publications, such as state universities and colleges several years ago. The practice changed and Robert tried to identify as many of those as we could to correct them to 'blank' instead of the 's' that was already there. We are not able to catch all of them because not all state universities and colleges are identified as such. So, there are records that still have the old practice. The Bibliographic Formats and Standards page on government publications does have the current standard, which is that state university and colleges publications should not be considered government publications.

To verify, local proxy 856s of provider links already present in the record seem inappropriate for the WorldCat record. Is this correct? We are seeing edit wars with deletes of some of these being detected and reinserted very quickly, sometimes in just a few hours.

Yes, they are inappropriate for the WorldCat record. We have seen those reappearing very quickly too. In most cases, we don't think they are edit wars. We think that they are coming in through our data sync service. We are working very hard to change the coding and the programming for our data sync service so that we won't import, or transfer, those local URLs into existing WorldCat records. We don't have an exact time frame on that, but we hope very soon to be able to prevent that from happening, in most cases. The thing to keep in mind with that is to do what makes sense for your local practice. If you take them out then replace the WorldCat record that is fine, just know that they may show up again a couple days later. That is not because someone else has put it in there, it is most likely because it was a batch process matching thing. We are working on that because we are just as frustrated about it as you are.

I often find DLC records that have 653 _0 fields that duplicate 6xx _0 fields. Since this is an uncontrolled index term, I don't understand why it contains an LCSH subject heading string. Was there some program run that created these fields? Can we delete them from the WorldCat records?

In looking at one of the records, it does seem like they duplicate the LC subject headings exactly, so I would recommend removing those. Without taking a look at the history of this record, I can't tell for sure how they ended up on this record, but most likely it transferred from ingest activity. Since the tag would be unique, different from the LC subject headings, they would be eligible to transfer. The transfer of field 653 is one of the transfers that we are trying to limit. If you are unable to remove those fields, you can report them to and we will take care of it.

How does P-N guidelines relate, or not relate, to microfilm reproductions?

They do not. There are no Provider-Neutral guidelines for cataloging microform reproductions. You still do those even if you are applying RDA rules, the same way that we did contrary to AACR2 rules, where you would describe the original, then give all of the publication details in a microfilm reproduction note. Including the fact that it is reels of microfilm, or that it is microfiche, or micro opaques, or whatever.

For Create Space publications, should the publisher be the author or Create Space?

This is a case for Create Space, it’s not so much Print on Demand reproduction as it is you just get a copy. An author goes to Create Space as the mechanism for publishing their work.  I am thinking that the Print on Demand model fits really well for something like Higginson Book Company that is in Salem Massachusetts, that reprints all sorts of old history books. Whenever you want one, they will produce one that is probably manufactured better than the original one was. They should be treated just like other Print on Demand publications, thinking of Create Space as the “publisher” because it is the author who sends their manuscript to Create Space and they do the work, not a lot of work, but of making it available. 

For POD publications that are a re-setting of the type, and therefore have very different pagination, does that count as a new manifestation and requires a new record?

Yes, it ought to have a new record in that case because it is not just solely the reproduction of what was there, it's basically a new manifestation.

I am fairly new to cataloging and I have a series of 16 books published from 1985-2018 that have different series names in field 830. Therefore, currently they are not all traceable (to one series name). Is there a way to edit the ARN record to include the additional series names so when a particular book is searched, it will list all other books in the series? There are 4 different series names and only two have the ARN number.

You could send an email to the email address. It would be great if you could send photocopies from the items that you are working with. We can create the series authority records that are lacking from the authority file for non-NACO libraries. It is hard to generalize a question without taking a look at what you actually have, whether or not it could be included on the existing series record or if we need to create new series records.

Why do you often see two records from a vendor (batch loaded), which are duplicates?

It would be great if you could send examples, when you come across these, so that we could examine why. If they are from the same vendor, it's probably a timing issue as to when they came in and they somehow didn't match up because of the timing in which they were being loaded. If they are from two vendors, perhaps there is something different about the records that caused them not to match. This is obviously a problem, and we want to know about them so that we can get them merged.

March 2019: Intro to the Virtual International Authority File (VIAF)

Topic-specific questions

Where exactly do you find the choice "All VIAF" in the drop down? I'm looking above the cataloging tab and the authority tab and I'm not finding that choice listed.

If you go to, it is on the top of the page,

Does VIAF have a mechanism for ingesting ORCID numbers?

Right now, if an ORCID number is in the 024, we do use it to match between source records. We don't ingest ORCID records at this time because the last time we tried it; the match rate was insufficient. The ORCID data that's in the public download doesn't enough dates and titles, so we weren't making good matches.

In Record Manager, I think you have the Dutch and German authority files available for catalogers? Is that part of VIAF?

Yes, there is a Dutch and German authority file available for catalogers in Record Manager. VIAF and Record Manager are receiving some of the same source data but they are not talking to each other and are processing the data differently. Therefore, what you see in Record Manager is not exactly the same as what you will see in VIAF.

Does VIAF have any rules about how geographic names must be formed?

We try not to have any rules about how names are formed and just use what's sent to us. Which means there is quite a bit of forgiveness in our matching algorithms because we realize that different national libraries have their own rules. We try to work with what we get. That's why those clustering numbers for geographics tend to be lower and tend to be what we call singletons (just means it's not clustered with anything else) because there are various rules that are used for geographics. They can be a little more complicated, as well as corporate names.

You said, "identifiers are persistent for the most part", what are cases when identifiers change and what is the percentage?

The clusters get rebuilt every week. The membership within the clusters as records are added and deleted, and occasionally because the algorithm has changed. So, if two clusters become merged, the one of the VIAF IDs lives on and one becomes abandoned. Although if you had a link to the one that's abandoned, it will always get you to where you need to be. We always redirect old links to where they should go. We make every effort to not have dead links, even though we do necessary maintenance on the clusters themselves. We calculated the percentage for a couple of specifics, so like LC and ISNI recently, and it was in the 99% range of the IDs in those sources were in the same cluster a year ago that they are today.

Is there any communication between OCLC and contributors when contributor’s data is causing algorithmic errors?

Yes. If there is any information that is inaccurate in the entities VIAF page, we get that information from our contributors, so we need them to fix it in order for the next data harvest to get that correction in.

What redirection happens for cluster splits? Does it redirect to just one?

If a cluster splits that has four member, and splits into two clusters with two members, two of them are going to keep the old ID and two of them are going to get a new ID. If you had a link to the old ID, it's going to take you to the old cluster that now only has two members. You can also save links to specific processed records, and that will always take you to the cluster containing that record. So if you particularly care where the LC record is or where the BnF record is, you can save a link to a particular BnF record and when you click on that link it takes you to the cluster containing that record always, regardless of where it's ended up.

Can you tell what kinds of traffic are coming into VIAF via the permalink ID? Is there evidence of the volume of automated calls on VIAF data?

Yes, and yes. Another one of our team members runs Google Analytics on VIAF, and he can tell who's been downloading things, who's been searching, and how many searches are from people, and how many searches look they are from bots. About 75% of traffic is coming via permalink IDs in VIAF. Our conservative estimate is that at least more than half of the transactions done with VIAF data are by bots.

Do any end-user systems make use of VIAF?

We're using the VIAF IDs. They are certainly forming our data link projects. They're going into the prototypes we've been doing for linked data projects. Jenny is also on the team that does the FRBR clustering, and the VIAF IDs definitely play a big difference in the way the FRBR clustering works.

What redirection happens for cluster splits? Does it redirect to just one?

In a split, there is really not a redirect. Some of the cluster members will keep the old ID and some of the cluster members will get a new one. So, there wouldn't be a redirect with a split.

How often does LC load to VIAF?


How quickly do staff respond to error reports?

We do get a large volume of VIAF requests. We do process them first come first serve. We do have a little bit of a backlog, but we are working through it. It's interesting just how many people use VIAF and actually, many are not in library-land. There's actually a lot of folks that edit Wikipedia articles that notify us with issues of clusters, as well as author's themselves are just general users. We will check them and get through them eventually.

Do you have any plans to work with subject authority files or other files (genre/form, etc.)?

Not at this time.

Does OCLC have any thoughts about the use of VIAF URIs in LCNAF authority 024s? Is it useful? Could/should the field be automatically populated based on LCCN presence in a VIAF cluster? 

VIAF will use that. If you put the VIAF ID in your 024, we use it for clustering. 024 fields that have different control numbers do help to create suggested links. So that's definitely helpful for our side in clustering. On the flip side being the person who uses that, if you don't maintain it and make sure it's right, it can get to be a problem eventually. If you make these links you got to curate them. It was mentioned that there is a moratorium on adding 024s in authority records right now for NACO.

Has DNLM headings taken the place of MeSH headings? I came across a record that had DNLM headings but it didn't have MeSH headings in the record.

A bibliographic record that has NLM as the source of the record and for some reason it does not have MeSH headings on it, would be very unusual. Normally, you would expect to see MeSH headings there, so 6xx fields with a second indicator 2, whether it's a 650 topic or a 655 form/genre heading with a second indicator 2. If you have a record like that, that looks like it's from NLM and is lacking those kinds of headings, send it to us so that we can investigate what's going on. If something stripped those headings off, we would be concerned and would want to take some action on it. If it was just issued that way, we could maybe figure that out.

(Laura) Maybe these are the headings that appear to be MeSH headings because they have second indicator 2 but did not appear to be coming from NLM? There was no DNLM indication in the subject heading itself.  If that is what the question is about, we are aware of it and cleaning them up.

(Robert) The record that was at the heart of this question appears to be a record that originally came from a vendor and then has been upgraded with a bunch of different symbols, the National Library of Medicine not being one of them. So even though it has medical subject headings that have been assigned, they look like they came from one of the libraries that contributed to this record rather than from the National Library of Medicine itself.

If PCC libraries start putting URIs in 024 rather than numerical identifiers, will you be able to use those?


If we report an error, do you send that information back to the original source institution/database?

If clusters get modified or changed, we are not letting or contributors or participants know because they're downloading the data. So those changes will eventually be found in the data file, which is actually updated monthly, unlike the live database which is updated weekly. So, they will get those changes. What we do report back out to our contributors is if there is incorrect data coming from their authority file or their bibliographic file. If a work is incorrectly associated with an author, we will notify the institution that is contributing that incorrect data because until they fix it, then it's not going to be reflected in VIAF.

If the only LC call number in a record is very general (for example NA680 for a non-general architecture book, or just PN for literature), can that be replaced with a more specific one? Or should we just add a second more specific one?

If you are able to replace the record and there is a legitimate, more specific classification, you should replace the existing one with the more specific one.

Do any digital assistants like Siri, Alexa, Google, Cortana, or Bixby use VIAF data?

We are not sure.

If we find a controlled heading in a bibliographic record that is not the correct entity and change it to the correct entity and control it, will that eventually trickle down to VIAF so that the work is not associated with the incorrect entity?

No, that data does not flow in that direction. The controlled headings in WorldCat affect the way Identities pages are built so if you control it to the right name, the Identities pages will be corrected. The works don't flow directly into VIAF. The works that you see on for entities are coming from bibliographic files that are sent to us from the VIAF participants. They are not works that come from WorldCat. The WorldCat Identities, which is another project, that data is being harvested from WorldCat. For VIAF, those works are coming from bibliographic files of our participants, and not all of our participants have records in WorldCat. Many of them don't, and many of them send their records directly to us.

Can you take advantage of 672 and 673 fields in authorities to help with matching? For clarification, the 672 and 673 fields have control numbers unlike the 670 field.

We don't attempt to do that now, but we will investigate it further. We do pull some data from 670. We look for some very specific format citations and patterns to find titles or to find birth dates from title pages. Sometimes there is a birth date that came from the title page of a book, and we have some real specific patterns that it can look for but that's about it. 

Are we allowed to use 386 fields in bib records? If yes, are they indexed in WorldCat?

Yes, it is part of the entity attributes and can be searched by en:.

What kinds of bibliographic records are sent to you as part of NACO?

Those are coming from the Library of Congress. The LC NACO file is a cooperative, there are many libraries, but the bibliographic records that are associated with any LC NACO records are just coming from LC and not from any other contributors.

Is any work being done to reduce the circularity between VIAF and ISNI? Sometimes badly-clustered VIAF data is used as a source in ISNI and then that ISNI source data reinforces the VIAF cluster. It becomes very hard to resolve errors.

We agree. At this point, we don't believe that ISNI has ingested VIAF data in probably a year. VIAF is continuing to ingest ISNI data. Otherwise, it goes back to once you make a link, you’ve got to nurture it and take care of it and make sure it's still good.

Does the Canada grouping include both English and French names for the same entities? With LAC contributing through NACO for English names, is there going to be duplication with LC and Canada?

When that whole transition was occurring, we did the work and now any LAC authorities that are English are in LC NACO and showing up there. The only LAC authorities that you would see in VIAF are from the French authority file.

Does OCLC have any plans to index the 008/24 Nature of Contents field in bib records in WorldCat? I am asking because it has the value 6 for comic books/graphic novels that can be hard to limit to in a WorldCat search.

We have no plans at this time but can certainly be looked at and investigate to see if it is something that we can take on.

LC is a contributor to VIAF and VIAS is a resource for LC authority creation. What future relationships do you anticipate between LCNAF, the OCLC authority file, and VIAF?

At this time, the relationship between all of us is working and don't see that changing at this point in time. Right now, VIAF is in a transition and it continues to change little by little, but at this point in time the relationship we have with LC and the other participants is pretty stable and don't see it changing.

Can you give specifics for the "weekly" re-clustering? Is it always the same day? I'm thinking of checking back to confirm a change has "worked", etc.

If everything works beautifully, the clustering finishes sometime Monday morning and it's visible by Tuesday morning. Occasionally there are issues and it's later in the week, but it starts on Saturday afternoon and it's usually done Monday morning and usually visible by Tuesday morning. Please note that if it was a request that you sent and it hasn't been processed yet by WorldCat Metadata Staff, then that change is not going to appear because that change needs to be done by staff manually.

April 2019: An overview of OCLC training resources

Topic-specific questions

What are the differences again between Connexion client and browser classes?

Differences between client and browser: Connexion client is software you can download from OCLC's website and install on your computer. It has more functionality for automating things, batch processing, etc.

Connexion browser is a website you can log into and you can do all of your cataloging on the website without installing any extra software on your computer.

So, the Connexion client classes cover the mechanics of how to use that interface and the Connexion browser classes cover the mechanics of how to use that interface.

Is the information in the WorldShare Record Manager policies different than what is in BFAS [Bibliographic Formats and Standards]?

No, it's not. In fact, we refer heavily to Bibliographic Formats and Standards in the policies classes, so it's really an amplification, an illustration of how to apply the policies in Bibliographic Formats and Standards.

What are the differences between Expert Community and Database enrichment?

Database enrichment has been around since 1991, enabling full mode and higher authorizations to add subject headings and call numbers to records. Additional capabilities were added to Database Enrichment in subsequent years, including fields 006 and 007, 300 physical description fields to CIP records, 505 contents notes, upgrading any CIP record except for its Encoding Level value of 8.

The Expert Community was created in 2009, expanding upon Database Enrichment by enabling full-mode and higher users to edit most fields in most records. The complete list of fields that may be added if not present, added even if already present, and edited if present to non-CONSER PCC records under the capabilities of the Expert Community can be found in Bibliographic Formats and Standards, Chapter 5.2, under the heading “Enriching PCC Records.”  If you try to add a field that is not on that list, then you will get an error message when you try to replace the record.

Is the OCLC training free?

Yes, all of the OCLC training is free. You do need to register for it, but there is no charge. The training we link to for other organizations may or may not be free. The Library of Congress training is free, but the training from ALCTS is not free.

Our DSpace repository is harvested via OAI-PMH. If I should change any of those records would the changed be overwritten next harvest?

Your DSpace repository is profiled by a collection manager at your institution using the WorldCat Digital Collection Gateway. Records are periodically harvested from DSpace and copied to WorldCat.

There are two factors that affect your ability to edit a record in WorldCat without fear of it being over-written by subsequent Gateway harvests. The first is in your control and the other is in the control of your collection manager.

If you edit a record in WorldCat and raise its encoding level above K, then the Gateway will never touch that record again even if it is enhanced or deleted from your repository. That’s in your control, but you’ll want to do that only in coordination with your collection manager.

The other mechanism is in the collection manager’s profile choices in the Gateway. The collection manager uses the instructions to Prepare a collection for the Gateway to set up the profile for your repository. One key setting that makes a difference to your edits is whether or not they’ve changed the default, Replace All, WorldCat Record Processing setting to Add New Only. If they’ve selected Add New Only, then you’re free to edit records that appear in WorldCat without fear of them being overwritten.

How do you qualify for Expert community?

If you have a full-level authorization or higher and can breathe, you are a participant in the Expert Community. No additional qualification is necessary.

Re: pcc records. Sometimes LC has started a record, but not finished it. Should we upgrade the record? It seems the records are anyway overlaid later if/when LC finishes the record. Also, sometimes their upload ends up resulting as a duplicate record if we have worked on their record.

Go ahead and upgrade the record as you need it to be complete for your catalog. There isn't a reason why you should not do that. It may end up causing a duplicate record to be added later if the data does not match exactly, but we will try to catch that using the various tools that we have, such as DDR.

A general question: Does OCLC currently do any ongoing maintenance of 776 fields? E.g., update $w OCNs for merged records?

For the longest time, we have not done anything with 776 fields, and this is true of many linking fields, where when a record is merged, the control number referenced in that linking field is not updated to point the current record. Of course, you can still search that number, but the number that you search may be in the 019 of that related record. It would be a useful thing to update the control numbers in these linking fields. We have started investigating that and will hopefully expand what we do so that we can get rid of duplicate 776 fields. Sometimes you'll notice on records that there will be two or more occurrences of a 776 field for an eBook version that point to different control numbers, but several eBook records have been merged into one. We only need one 776 field in that case rather than 3. So, we're looking into ways to get this data up-to-date.

I would like to see training on the Connexion Digital Import function or if there is something in WorldShare Record Manager for uploading to CONTENTdm.

Thank you for that feedback. We will take that into consideration. 

Are there comparable options to edit WorldCat records for non-English cataloging communities?

For those who catalog in languages other than English, the Expert Community and all of the other--minimal-level upgrade, database enrichment--they apply to those records as well, so someone who is cataloging in a non-English language, they can edit records in their language--there's no separate program for that.

Are there non-English versions of the training videos, etc.?

To a limited extent, yes. For example, if you go to the WorldShare Record Manager training page, at the bottom, there is a link that says, "For WorldShare Record Manager training in French, please click here." You will find some videos about Record Manager. We don't have a lot of non-English training content, but we are working to get more.

How crucial is it to update things like access point forms in linking fields? Example: the NAF form of an author's name changes, when doing BFM, update all linking fields as well? For future 776 fields updates, is it possible to also update the $a when the authorized form of a name has changed?

It would be really nice if those were all updated, but it's not crucial that they all are. Hopefully, down the line, ideally OCLC would make use of the control numbers in pointing to another record and make use of the data in that record, rather than having to manually entering data into a linking field and then having to separately maintain it. We have done some maintenance on headings in linking fields in the past. For instance, when we have gone through the database to update AACR2 forms to be compliant with RDA, so if we have an author where it says "died", we were going to change that to a hyphen and a date. We did that in linking fields as well. But routine maintenance of those where the heading changes, doesn't happen unless catalogers go in and manually maintain it. It's just a citation, so if it's a little bit off, it's not the most critical kind of thing, but it's up to catalogers to maintain those.

Bibchange does receive requests for BFM, to update these headings, and when QC staff works on that, they also update linking fields to reflect the current form. So, you could submit a request to and we will do that for you.

Since April 8 there is a new pcc-policy concerning the end-period. What are the implications for non-pcc libraries? Are they supposed to continue the former policy?

At this point, you could follow PCC practice and omit the periods, which actually is the easier thing to do, because if you are familiar with MARC 21 documentation, and what we've also included in BFAS,  some fields traditionally had a period added on to the end and other ones do not. It's kind of hard to remember which ones fall into what category, so it's generally easier to not include the period. However, at this point, you can continue with the practice of entering the periods and that's not a problem.

Could OCLC give training/instructions on working with very complex searches (in RecMgr)? When using many different indexes (for both bib records and LHRs) and operators, we often get unexpected results: too many or too few.

We do have 2 videos on searching in Record Manager, Basic and Advanced Searching. Having said that, those videos are probably not going to address your specific questions about very complex searches. If you have a search that isn't giving you the results you want, send your question to, or call the toll-free number, 1-800-848-5800, and someone can help you with that search.

What does leaving out the end periods do with the presentation in WMS and Discovery?

It doesn't have much of an impact because the bibliographic data in so many of these displays is taken from the MARC record and labelled, so there's a caption and the data, the lack of a period at the end is probably not going to affect somebody's ability to understand what they are seeing.

I would appreciate training on WebDewey and its number building feature.

We have some training on WebDewey, although we have not offered a training course in the recent past. An older introduction to WebDewey is available at In addition, ALCTS has a video on WebDewey Number Building Tips and Tricks at

When PCC refers to terminal punctuation are they referring only to periods, or to all ISBD punctuation between subfields?

The announcement that was made recently is only referring to terminal periods, the punctuation that is at the end of the entire field, not the ISBD punctuation that separates subfields.

When reporting duplicate bibliographic records, I often see duplicates from national libraries for the same bibliographic entity. I am not sure if I should be reporting these. Is each national library allowed to retain its record in WorldCat? For example, if there is a record from a U.S. member library and duplicates from the Australian and Canadian national libraries, which records get retained or merged?

There should only be 1 bibliographic record representing a particular resource, per language of cataloging. There are certain situations, a hierarchy, for which record would be retained, based on the level of completeness.

PCC members are no[w] being asked to report NACO activity to LC. Currently there is no differentiation in OCLCs stats program for NACO between name and series authorities. Is there any chance this distinction could be added to the way OCLC tracks and reports NACO work?

That would be a useful enhancement to the participants, so we can put it on a list for future consideration.

Regarding the new PCC policy regarding limited use of ISBD punctuation in bibliographic records (April 8, 2019) PCC libraries may stop terminal periods. What should the Desc in the Fixed Fields be, if we begin excluding the terminal periods? For non-PCC libraries, do you recommend that we also begin excluding the terminal periods?

You would code Description in the Fixed Field as you always have, because the other punctuation remains in the record, so it's only terminal periods that you would potentially exclude. For non-PCC libraries, you may include or exclude terminal periods at this point, as you see fit.

For more information, see the PCC's Guidelines for Implementing New PCC Policy Regarding Terminal Periods in MARC Bibliographic Records.

Are there plans to add other subject authority files for automatic verification? like lcgft? or others? [Including] 650s with 2nd indicator 7 and $2. For example, being able to control headings from AAT

For those that are not currently available, we are open to exploring options to increase the number of thesauri available. Verification and controlling would be limited to Record Manager.

Is there any foreign language authority file added to OCLC (for example, any Spanish language naf)?

We currently load the German and Dutch name authority files, along with French names from the Library and Archives Canada, as well as Maori subject headings. These are all available in Record Manager, but not Connexion. There are no immediate plans for adding Spanish authorities, but we are open to exploring that possibility.

What is the state of migration to Record Manager? How much longer Connexion will survive?

There is no set timeframe for the end-of-life of Connexion. If you catalog using Connexion, you are free to switch to Record Manager. Some of the functionality available in Connexion does not yet exist in Record Manager. On the other hand, Record Manager has some features that don't exist in Connexion. So, it's a matter of what works best for you, and that may be all we can really say at this point.

I found the concept of a named conference a little bit difficult can you explain?

For example, if you are cataloging the papers or proceedings of a conference, normally, the conference will be named on the source with an explicit name, as opposed to, well, we had a meeting where this was discussed. In the latter case, there's no formal name for the meeting.

The Library of Congress' refresher training on Conferences (updated October 2012) states:

A named conference need only fulfill these RDA definitions

  • RDA 8.1.2

Person, Family, and Corporate Body:

“The term corporate body refers to an organization or group of persons and/or organizations that is identified by a particular name and that acts, or may act, as a unit.”

  • RDA 8.1.3


“The term name refers to a word, character, or group of words and/or characters by which a person, family, or corporate body is known.”

Can one use both Connexion and Record Manager?


Is Record Manager available for NACO?

Yes, NACO institutions are currently able to use Record Manager for creating or updating name authority records. This functionality has been available since September 29, 2018.

Is there somewhere where the differences between Connexion and Record Manager are listed?

We are not aware of a specific list of the differences. Record Manager is constantly being further developed, so it would be a constantly changing list. This may be something we would want to bring up with the people involved in Record Manager development, to help people decide whether they want to move their work.

Would the ALA annual conference be a "non-example" of a named conference? I.e., it would be a 110 rather than 111?

You could have conference entered subordinately to a corporate body, but it would still have a name.

Is Record Manager only online?


Is there any way (aside from looking at each record) to distinguish which record have controlled vs uncontrolled headings for the same name?

There is an index for searching controlled names, "ar:", followed by an ARN. So, for example, if you wanted English language-of-cataloging records containing "Lee, Stan" that are not controlled to "Lee, Stan, ǂd 1922-2018 (LCCN n 83169267; ARN oca00990699)", you could enter this Command Line search: na=lee, stan AND ll:eng NOT ar:990699.

What is the status of FAST headings? Are FAST headings still considered experimental?

We are in the process of moving FAST into a production mode as opposed to an experimental system that is centered in our Research area. So, they are not considered experimental anymore. The production mode is coming soon.

Will FAST headings only be in English?

Yes, as things stand right now, because Library of Congress Subject Headings have been the basis of FAST headings and those are in English.

Can members add FAST headings? or you have to wait for the automatic processing?

You can assign FAST headings to a record for an item you are cataloging, or if you find a record in the database.

May 2019: Small errors with big consequences

Topic-specific questions

Are mixtures of Latin and non-Latin characters usually human introduced, for example, someone with a different keyboard encoding manually typing one or a few characters in a field, or...? What is the best prevention?

It is very likely that they would be human-introduced especially if the cataloger wasn't typing on a multilingual keyboard or wasn't used to entering text in a particular language. On the other hand, the cataloger may have copied characters from a Web page and pasted them into the record.

There is also the scenario that script, such as Cyrillic was entered by the cataloger but the cataloger forgot a letter. Instead of adding the correct Cyrillic letter, though, the cataloger added the Latin form of the letter. This is similar to the transition from typewriters to computers, where the letter "l" was no longer used interchangeably with the number "1". When using computers, these are different characters. The same thing is true with the Cyrillic vs. Latin letters. They are different characters and should not be used interchangeably.

Sometimes pasting summaries or abstracts into the 520 introduces character errors.

The problem is most likely the smart characters when picking up information from a webpage and pasting it in a record.

How do you prevent the smart characters from getting into the record?

Convert the smart characters before pasting them into the record. One way to do this is using find and replace within a Word processor. Another way would be to create a macro to convert all of the characters automatically. Macros can be created in both Connexion and other programs such as Notepad++.  There are also Connexion macros available in repositories such as GitHub. An example provided by a library in the text chat suggested PasteUnformattedPlus, which can be used when copying text from other sources and can be downloaded from

The 'Cataloguing defensively' instruction tells us not to correct minor errors. Do the examples given count as major errors or are we now encouraged instead to do these minor corrections?

You are welcome and encouraged to correct errors in cataloging if you wish. This is a hallmark of the cooperative cataloging environment in WorldCat. We will review our instructions on cataloging defensively to see if any corrections need to be made.

It may also be the case that, the statement to not correct minor errors in the cataloging defensively instruction was a cautionary statement. If the situation is that the difference is minor, then do not change the WorldCat record. For example, if the pagination in the bibliographic records is "330 pages" but the resource you have in hand is "331 page" then use that WorldCat record but only edit your local copy if you wish to change the pagination. These would be considered copies of the same edition.

However, the small functional errors discussed in this presentation, like a comma missing from a personal name, can have an effect on that name being found, or if the non-filing indicator is incorrect, that has an impact on that title being found. These are the minor errors that are worth fixing.

I find a number of mistakes with the FAST headings. Is it a good idea to just delete FAST headings?

The answer depends on the context. If the FAST heading is inappropriate for the item, then since they were generated from the LCSH, you are encouraged to fix both the FAST and LCSH in the WorldCat record. If the error involves a FAST heading that was mapped incorrectly from the LCSH, then report these errors to so that staff can fix the mapping to prevent the problem from being repeated on other records. If the FAST heading was manually added to the bibliographic record and a typo was introduced, then you are welcome to correct the heading or report it to for staff to review and correct.

How often do you do this correction work? Is it constant or project-based?

For the things previously mentioned, we sometimes rely on people spotting and fixing the error themselves when possible. There are certain errors that we can go looking for because they are readily found. In some of these cases, the number of occurrences is overwhelming so that we are constantly trying to catch up.

One example of this is the incorrect non-filing indicator. Entering the correct non-filing indicator often does involve a detailed knowledge of the language of the resource. It is not as simple as setting the non-filling indicators to a certain value if you see a particular word. For example, the word "the" in an initial article in English but if there are diacritics such as in Vietnamese or French, then it should not be coded as an initial article as these are different words altogether. Another example is the letter "A" in English. Titles for children's works might begin with "A is for... ". In this case, we want to file on that "A" since this is talking about the letter "A" and not the initial article.

Some of these errors, while small, do take a lot of effort to correct, although some so lend themselves to macro corrections. For example, spacing problems related to the comma used in a personal name access point. This would affect how the name would appear in a browse list. It is difficult to insert a comma into a personal name that is missing one. To insert the comma, you need to determine where the surname ends and the forename begins as well as how many words the surname includes.

In your slides you showed 'WorldCat browse list', for titles, for Personal names. We use Record Manager and haven't found such functionality. Does it exist in RM? We could really do with such browse list in order to check Authorized Access Points!

While there is browse functionality in Record Manager, it is not exactly the same. A library suggested in chat that one way to browse in Record Manager is to change the scope to "All WorldCat Browse" then use an index such as "Personal Name - Whole Phrase". More information about searching using browse in Record Manager can be found on the WorldShare Record Manager Search page. If you have further questions about the functionality, please contact OCLC Support.

I see a lot of subject headings coded as LCSH when they are not valid LCSH headings and more likely something used locally. What is the best way to handle this situation?

You can re-code any non-LCSH to show that it is a local subject heading by using second indicator 4. You are also welcome to report it to so staff can investigate it further especially if you are seeing a pattern from a specific institution.

Going back to correcting minor errors, sometimes I have wondered if some corrections were based on opinions rather than correct vs incorrect. For example: For a week, I got the same record with 100 field corrections DR. then Dr then DR. again and back and forth.

The name should be based on the author's usage and you would normally apply rules for capitalization, which would mean that "Dr." would be the chosen form. You are welcome to send this particular record number to and staff can take a look to see what is happening with this record.

Will OCLC allow 020 added for reprint periodicals? Here is the email I got from Les Hawkins, Coordinator of CONSER: "publishers will frequently issue an ISBN for a complete reprint, possibly of a ceased serial or as your example shows an ISBN for each of the 6 volumes of a ceased serial., This might be a good case for OCLC changing the validation of the 020 fields if there is a way to distinguish when the someone is recording ISBN for reprints vs. when they are recording ISBN for each issue of a (non-reprint) serial."

This is an issue that has come up now and then. Back in the days of AACR2, instructions in Chapter 12 told catalogers to record an ISSN in the record but there was nothing that said to record an ISBN even though we treated serial reprints that came out all at once as serials and cataloged them according to Chapter 12. Reprints of serials will sometimes have ISBNs printed on them and so they do seem valid to include them. In RDA, instructions tell catalogers to record standard numbers, which leads to the question of including ISBNs in serial records. This situation was brought up a few months back on the CONSER list. While adding ISBNs to serial records would be nice to do, it is technically difficult for us to do at this point. In the MARC environment, you don't want to have the situation where all of the ISBNs relating to individual issues are listed in separate 020 fields. If you did this, you might end up with 100's of 020 fields in a serial MARC record, with each 13 digit version being duplicated by the 10 digit version. The consensus in CONSER was that would be nice to do for reprints at least, but not at this time. When we move off of the MARC format, this would be much easier to implement. For the time being, we suggest that you add these ISBNs to your local copy of the record.

Are AU@ Libraries Australia MARC records merged with other eng cataloging rules MARC records (say PCC or DLC) or are AU@ MARC records retained as parallel records?

All records cataloged in English are merged to any other records cataloged in English if they are duplicates, including AU@ record.

A record for Stephen King's original Pet Sematary has the ISBNs for multiple editions, including the new movie tie-in. This was very confusing. The Stephen King Pet Sematary is #14046555 or #9576009. published in 1983. None of the incorrect ones are in subfield z.

The ISBNs could have been matched to an existing record in WorldCat representing the book from some years back, however, the ISBNs for the newer editions and the movie tie-ins don't belong on the WorldCat record. For ISBNs, you would record all ISBNs that appear on the item in hand in field 020. The ISBN representing the resource itself would be recorded in field 020, subfield $a and ISBNs representing related resources would be recorded in field 020, subfield $z. If ISBNs don't apply to the item that the record describes, such as a movie tie-in or a later edition, and they don't appear on the item that the record describes, then those ISBNs should not be added to the WorldCat record. For this particular example, we've cleaned up the records you've mentioned along with some related records based on available online proof and evidence found in OCLC's Journal History.

If you do come across errors that you are unsure whether to submit or correct, please feel free to reach out to or and staff would be happy to investigate further. If a large problem is identified, then staff can perform scans to compile the records for correction.

I'm cleaning up some old records and matching them to OCLC records. I find a number of these older titles have multiple mergers in the 019, and there are fields for a narrator on a print record, a note for the artistic qualities of the illustrations for a presumably special edition.

This is a good example of errors to report to for further investigation. Staff will be able to determine if an incorrect merge occurred or if there is something else at play.

I semi-regularly see ISBNs on records for early books that couldn't have ISBNs. 

It is possible that there was a reprint that had an ISBN that was added later. In this case, it would be appropriate to keep this ISBN on the WorldCat records since it represents the same manifestation.

In Record Manager, there is a filter to search only OCLC Publications New Zealand records in Advanced Search. Is this filter based on the holdings of NZ or by some codes in the records?

The filter option for "OCLC Publications New Zealand records" in Record Manager limits the search to records with the code "nznb" in field 042. This would retrieve results similar to if you limited a search by authentication code in Connexion, FirstSearch, or Discovery.  

June 2019: Merging duplicate bib records and the Member Merge Project

Topic-specific questions

Query - the LCSH heading that was carried over had Fiction not Juvenile fiction. This means that I would have to make a correction before accepting this record.

Headings that contain incorrect information, such as subfields, can be corrected either prior to the merge or after the merge is complete. Metadata Quality staff do try to clean-up the records using a clean-up macro and make any necessary corrections that they find as they are merging duplicates.

How does the clean-up macro work? Does it look at the codes and change the fields, or look at the fields and change the codes?

It is a little more complicated than that. The internal macro that we use has more than 30,000 lines of code, so it is very complicated. It looks at many, many factors trying to clean up obvious issues and problems in the record. Internally, we call this our QC or Quality Control macro. It works just like any other macro, it's just longer and more complicated.

How many members are merging serial records?

At this time, the University of Oregon and UC San Diego are the only institutions currently merging serial records, with another institution in training.

My notes from an AskQC session about DDR say that the 300 field could have two subfield $a's such as 1 score ; $c 28 cm + $a 4 parts. Is it okay to repeat subfield $a in the 300 field or should it be subfield $e? At the time, DDR (Duplicate Detection and Resolution) was ignoring subfield $e and this caused records with CD + book to be merged incorrectly.

DDR does take into account accompanying material, however that information should be recorded in subfield $e of the 300 field, not subfield $a. If a record has accompanying material recorded in 300 subfield $e, it will not get merged into a record that does not have accompanying material.  At one time, DDR used to handle this differently.  We changed this a few years ago so that accompanying material is taken into account in the Duplicate Detection and Resolution software merging process.

Since this is for PCC only, is there a path for those of us who are not to request a merge from a participant, rather than going through bibchange?

No, there is not. We still ask that you report duplicates to bibchange, if you are not a Member Merge Project participant.

Does NACO funnel participation count as PCC participation?

Yes, if your institution is a NACO funnel participant, you may contact us to participate in the Member Merge Project.

With this new program, how long on average does it take to process a merge request?

We don't track how long it takes to resolve an individual request. We have requests that come in for different formats that are addressed by different staff members, so the turn-around time can vary for different formats. We have some formats that are up-to-date and we have others that have several months of backlog. 

Is merging happening more quickly now that some members can merge as well?

The majority of the members who are merging are not merging from the backlog here at OCLC but are merging things that they come across in their daily workflow. While it may be making things faster because they are coming across less duplicates as a result, it is not diminishing the backlog that has been submitted to OCLC, at least not at this time. 

How long does it take to merge records when I send a merge suggestion to

That does depend on the format. We do have different turnaround times depending on the format and the number of staff available to review those requests. The books format backlog is our most extensive backlog at close to a year or more at this point. In general, without breaking it out by format, our backlog consists of approximately 40,000 requests to give a sense of how much we have to work on.

We use Record Manger, if we wanted to participate, would you have protocols, scripts, or tools for Record Manager or is Connexion used?

At this time, it is a Connexion only function. We don't have it built into Record Manager.

How do we know when to report merging records when one has been accustomed to seeing so many records for a particular manifestation (incomplete records, half-finished records, etc.)?

If you believe that they are duplicates when you look at them, feel free to report them. We can make the determination when reviewing them if they are indeed duplicates and merge them, or if not and leave them as separate records.

My institution has done a reclamation project which created a lot of duplicates in OCLC. Could this have been prevented on our end or yours? Most of our records had old OCLC numbers on them.

My institution has done a reclamation project which created a lot of duplicates in OCLC. Could this have been prevented on our end or yours? Most of our records had old OCLC numbers on them.

Does field 035 (System Control Number) with subfield $z have something to do with merging?

No, not that we are aware of.

Is there a quota for how many records an institution should merge per year?

No, we did not set a quota. It is really dependent on the institution and what they are able to fit into their workflow. 

If I report a record that needs to be merged, do all member libraries get that message, or only OCLC?

Only OCLC get merge requests at this time.

Do you pass the requests submitted on the web to the libraries in the project, since you have 40K+ requests?

We would love to, but we have not had that request from any of our participating institutions. When first starting in the project, some institutions don't have duplicates put back that they've come across in their daily workflow, so we will use requests from the backlog for training purposes. Once the institution becomes independent, their time spent merging is for those that they come across in their day-to-day work.

I rarely find duplicates for the books I catalog, but I find quite a few when I catalog scores. How would this play out for training purposes?

There is training documentation for merging scores duplicate records. So, if you only work with scores material, you may request to become a participant in the project to get training to merge only scores duplicates. Initially when we started this  project, we had the requirement that all institutions went through the books format training. We then realized with other institutions requesting to join that didn't catalog books, it didn't make sense to have that requirement. So, we are working with a few institutions that are training on the specific format they are interested in.

When we merge records, we have to transfer Google links. Two questions about this related to serial merges, which can have several hundred Google links to move. 1. Is there a macro that can help this transfer? 2. Is there any thought going into not moving all these links in the future?

When you do a merge, those will automatically transfer during the merge process. We are in discussions with changing the Google process, and all of those Google links is something that we have talked about improving.

Do Google links transfer automatically into CONSER records?

Yes, they do.

I have a list of a large number of records (about 80) that I believe were improperly merged. They each have an 019 field and inconsistencies in the formats across the fixed fields, 006/007, and 33x fields. What would be the best way to send such a large number of these at once? I've been simply typing up the details in a Word document, is that something that you could work from?

Yes. Record numbers for those records that you believe to have been incorrectly merged can be put in either a text file or directly in the email message and submit them to bibchange. We will review them on an individual basis to determine what happened and recover, if necessary.

I remember the 1st cohort required a massive time commitment from OCLC staff, so much so that expanding the project was in serious jeopardy. Now, 14 more institutions have been added, and other PCC institutions are encouraged to join the project. What changed to make it possible for so many more institutions to participate?

Several things probably played into that - a learning curve, being one. We were starting up something new with the 1st cohort and we really didn't know what to expect from that. We have continued, over the years, to improve and tweak the program. We also almost doubled our number of staff that are reviewing, so we were able to take on more institutions. Using an external review also helped with that process. We realize that this is really important to our member. It does seem like a big time commitment for OCLC but in the end, it really does pay off. As you saw from the number of records that the members are merging, in the end it's a win-win.

Do you have numbers of members participating for the different formats (for example, scores) and the number of score records merged?

We get the OCNs for all the merges that are done, but we don't have any way to determine what the format is for that. Those participating in the project can go to the OCLC Usage Statistics and see the numbers that they have merged, but that is all the information that we have available from our stats.

If I send a merge suggestion to, will I get any feedback about what happened? When I send a suggestion, if they don't get merged, I would like to know why.

With the number of duplicate requests that we receive in a day (over 100 requests), there is no way to provide feedback with the result of the request while trying to process as many requests as we possibly can. We are in the process of streamlining merge requests to process them quicker. This process strips out the contact info to harvest the OCNs reported, leaving no way to provide feedback afterwards. For reports of believed incorrect merges, we do provide feedback as to whether the merge was actually correct because the merged record contained incorrect fields that transferred during the merge process, or that the merge was in fact incorrect and the record(s) has been recovered.

I've been working on matching our old WLN records to the OCLC version, and I've found a number of records that, based on the variety of notes (5xx) are mergers of special editions (e.g., artistic illustrated editions, etc.) that would seem to have been better on their own record.

If you believe that records have been incorrectly merged, please report those to bibchange so that they can be investigated and pulled apart if necessary.

Does field 019 have something to do with merging records? Sometimes I take the OCLC 035 field and search it in OCLC and it appears in the 019 field and not on top where the OCLC number is and is really confusing.

The 019 field contains the OCNs of the records that have been merged into that record. The OCLC number that is found in the 035 field of your institution's record could appear in the 019 field of the WorldCat record in WorldCat, showing that record has been merged with the WorldCat record in WorldCat. The reason for retaining these OCNs in the 019 field is so that if your institution does have the OCLC number for one of those and search on it, it is still indexed and you can find the record that it has been merged to. So, it is very important to retain these so that the library can find the remaining record.

Besides the 019 field, are there additional tags to identify merged records?

No. The only way to identify a record has been merged is by the presence of that records OCN in the 019 field.

Are the 33x fields taken into account when merging?

Yes. The 33x fields contribute to the formulation of the material type and are taken into consideration for DDR (Duplicate Detection and Resolution) and other automated processes. For manual merging, they are given less consideration because we see the material type intended in the record and can determine if the 33x fields are correct or not.

Can participating institutions merge across formats? So, if one record is for a CD plus text and another for the same item is for a text plus CD?

Yes, these could be merged, but there could also be a good reason to leave them as separate records in WorldCat. It depends on the context, and you would want to consider each case individually. If one was poorly cataloged and the other is much better cataloged and has a lot of holdings on it, that might be a reason to merge them. If both are being used equally and the cataloging institution had a legitimate reason to place one as the predominate material over the other and some other institution made the opposite decision, those could continue to be separate in WorldCat. Bibliographic Formats and Standards chapter 4, When to Input a New Record states that if you do encounter a record where you disagree with the choice of predominate material, to use the record rather than creating another one.

Where do I find the number of records I have merged within OCLC Usage Statistics?

When you log into the Usage Statistics, it will be the second-to-last category at the bottom that says Merged Records. When you open that, you will see Institution Monthly Merged Records. Within that, you can enter your institution's symbol to retrieve the records merged for the given range. Please reach out if you are having trouble accessing your stats.

Are you able to merge cataloging records with non-English language of cataloging?

Records with the same language of cataloging (i.e., both records are German, English, Spanish, etc. language of cataloging) can be considered candidates for merging if their descriptions match for the resource described.

What if the language of cataloging seems to have been in error?

We ask that you report those to OCLC for further investigation. We can check the history of the record to see how it came into us and if it was possibly incorrectly changed by another institution. The revision of Bibliographic Formats and Standards chapter 2 will include an updated section for language of cataloging records and how to handle hybrid records.

One of the problems I've seen with using Data Sync to merge records is that libraries can do what they want locally with subject headings, etc. and I see a lot of bad headings coming in that won't pass OCLC's validation process.

So, this is a comment about OCLC's Data Sync loading.  This is not about Member Merge or people manually making decisions about what to merge. This is about how our matching software, in the loading process, matches the records and there are fields that are transferred from the incoming record to the existing record in WorldCat. We have just recently made some adjustments to that field transfer process, that hopefully will improve things so that not as many things that are perceived as bad headings will be merged.

We've come across records for online materials with no 856 fields. Under what circumstances are these made and can we add 856 fields to these records?

Yes, please add 856 fields to those records. Sometimes we are getting records from vendors, where they don't have a generic 856 field to supply because that vendor generates only individual ones for an institution. Ideally what we want in WorldCat is a much more generic 856 that anyone could click on to at least get to a sign-on page for the provider.

Do you know how many enhance/replace transactions by a member library are done and what years of publication generally are included?

We do have statistics on enhancements and replace by member libraries, but we do not have it broken down by years of publication.

Does OCLC create a 050 field from a 090 field during quality control editing?

Yes, the QC macro that we run on records does change the 090 to 050, if not already present in the record.

How would you be able to tell if a record was derived from another.

There are various areas within a record that could be a clue that the record was derived from another record.

Couldn't the derived record be in error? Creating a new record when it was not justified?

It depends on the records. You have to look at the metadata, such as if things are coded incorrectly or if there is cataloger supplied information such as an edition statement to help indicate that it was intended to be a unique record. If the records do not have that differentiating information, then it will be hard for a program to make that judgment, let alone a human, to make the determination that they are in fact duplicates or are meant to represent different resources.

Why does OCLC allow subject headings and genre terms in multiple languages within one record when the language of description is only for one language? In our library consortium, we delete all non-English subject headings, and we have to spend effort removing non-English after importing (particularly for the large number of French and German headings in those records).

The language of cataloging code found in the 040 subfield $b is specifically about the language of the description, as correctly stated in the question. So, that doesn't say anything about what kind of subject headings that can be on the record. OCLC allow these because we are membership organization, and there are member libraries that are out there that may want French language subject headings on records that they catalog in English, or may want Spanish language subject headings for records they catalog in English, or English language subject headings for records they catalog in Spanish. We don't want to impose a policy on member libraries that would preclude that. That is why you will see a huge mixture of subject headings on records of all types of languages of cataloging.

When we are upgrading a CIP record, why are we unable to change the encoding level from pre-publication (8) to full-level?

There are multiple reasons. To change the encoding level, an institution would need to be able to create authority records if they are adding subject headings or access points, particularly name access points to the record. Those records, especially those from LC, are PCC records and need authority records to back up the heading. PCC libraries can do that upgrade and change the encoding level. We also have some CIP upgrade program participants that are specially authorized to do that.

I've compiled a list of institutions that add institution-specific URLs to 856 (epoxy addresses). I think they are editing them through WMS. Can I send this to you? I recently replaced a lot of old records with newer ones with a WorldCat query. It added a lot of bad links to our catalog.

Yes. One of the things that we do when we are working on records for various reasons, is trying to delete local URLs when noticed. We realize that those clutter up WorldCat, so if you have noticed a pattern, please send those along to us.

August 2019: OCLC cataloging policies: An overview of Bibliographic Formats and Standards

Topic-specific questions

I'm not familiar with the "Functions" column in a Connexion record that was shown in the demonstration.

Going back to the slide of the Connexion browser, over on the right you can see a “functions” drop-down menu. You can access help from that menu. You can also select to add multiple fields or copy a particular field.

How do I know when a page in BFAS has been updated?

There is a date at the bottom of each page that shows when the page was revised in some way. It may be to just fix a typo, add more information or completely revise the whole page. 

Does BFAS indicate if/when OCLC will remove a field through its QA processes?

Once obsolete fields have been removed from WorldCat, the same information will eventually be removed from Bibliographic Format and Standards (BFAS).

Could you talk a little bit about where (if?) BFAS differs from MARC21 Format for Bibliographic Data and if there are differences, how decisions are made to diverge from MARC21.

We try to follow MARC 21 as closely as possible. If OCLC has added something in advance that is specifically OCLC defined in the past, we reconsider the element and decide whether we should get rid of it. In some cases, we have implemented content designation in advance of MARC 21, or if it comes to be the case that MARC 21 now has a field or subfield available to record the same kind of information, we might get rid of this to be in line with the MARC 21 standard. As part of that, we would convert data in the database or possibly eliminate it to take the obsolete field out of the validation and out of the BFAS documentation.

Often, our librarians print out a document found on the internet and want it added to our collection.  Would we use the section 3.1.2 and treat similar to a photocopy?

Yes, chapter 3.1.2 says “A photocopy or POD reproduction may be made in-house, or it may be ordered from a photocopy or POD service provider. The source of a photocopy or POD reproduction may itself be a reproduction.” Read the OCLC policy that follows this definition and treat as an on-demand publication.

Should we be reporting all the level M duplicate records? Is an ISBN that brings up multiple records enough?

Absolutely. Report any records you feel are duplicates in any way that is convenient to you.

I am a new cataloger, and wonder if Connexion browser is still available if our institution only pays for CatExpress ?

Please reach out to OCLC Support and they will be able to help you.

I see there will be a November session comparing Connexion to Record Manager. Is this an indication we should be learning Record Manager if we currently only use Connexion?

It would not hurt to learn to use Record Manager, but no, this is not an indication you should be learning Record Manager. There is no end-of-life date for Connexion yet, but if/when there is, there will be plenty of notice for everyone to make the transition.

Could you say something about Encoding levels (ELvl), recently I edited the record with ELvl 1 and cleaned it up and I did not change the ELvl, but I think that I should have changed it.

It sounds like the question is about upgrading a record. The information that you can provide and the types of changes that you are able to make to records can be found in BFAS Chapter 5, Member Capabilities.

Could you talk a little bit on reporting an error, for example for field 505?

How to report errors is covered in BFAS Chapter 5.5, Requesting Changes to Records. This also includes the reporting methods available.

Any chance that OCLC staff, perhaps some from QC, might begin similar standards for KBART Metadata. With budgets moving from individual print books to purchasing collections of eBooks, quality control for KBART data is needed. It feels like there need to be a partial shift of resources towards the quality of collection metadata. Any comments? 

There is a session coming up on Knowledge base quality that should address this question. We will be sure the person giving the session will be given this question because it may help with their presentation.

Should we report information missing in the 300 field, such as illustrations, and the pages in 504?

We do get a lot of requests for this. If this is an edit you are unable to make, but it is important to you, please send it to

Re: FAST headings:  If we add, delete, or change LCSH in a bib record, we should delete existing FAST headings, is this correct?

Yes, if you delete these because they are inappropriate or incorrect, they should get regenerated correctly. If you are adding another additional LCSH headings, then we will go ahead and add the new fast heading at a later date.

I find that BFAS has better examples than MARC21 under each MARC field. The examples in BFAS are broken out by subfield (i.e. you can find examples for every subfield). It's also easier to see the correct punctuation between various subfields.

Thank you, we worked very hard on that. We try to provide as many examples as possible (and stewed over the examples in some cases to illustrate as many situations and subfields as we could). If you happen to come across any good examples that are not already exemplified in BFAS, you are welcome to send them to us to help provide better explanations.

At one time, I seem to recall hearing something about Connexion changing to a web-based interface.  Are there currently any plans to do this?

We do have the Connexion browser interface that is web-based and Record Manager WMS is web-based. Only the Connexion client is a desktop application. If you would like to try Record Manager, you can sign up for an OCLC Services account, see what you think of it.

Does OCLC have any plans for additional vocabularies for the control headings function, in particular the Getty Art and Architecture Thesaurus?

Heading control functionality is added with new authority files, but that functionality exists only in Record Manager rather than in Connexion. I believe that Getty Art and Architecture Thesaurus is one that is on the schedule to provide the control functionality.

How do I report errors in the text of BFAS?

Send error reports to But of course, there are no errors in BFAS!

Record manager doesn't have macro capability for record clean-up yet, does it?

Not yet, but they are working on bulk edit for local records, but not yet on WorldCat records. However, there are some advanced actions available in Record Manager, such as deriving an electronic record from a print record.

In RM, the bulk edit for LBD and LHR (pilot) uses scripts, not macros. The LHR pilot will run from end of Aug to Sept.

Why do some subject headings in the authority file not control when added to the record? Example: Imaginary creatures $v Fiction.

That may have something to do with the type of authority record it is.

Most often the case is explained in the 667 field, which may say "Record generated for validation purposes" and is generated off the LC bibliographic file as opposed to intentionally put into the subject authority file. Many are machine generated and are not used in the controlling process because many consists of a topic and a free-floating subdivision for which there are other authority records that can possibly be controlled incorrectly. We opted to exclude those that have the 667 note that says it was automatically created. We rely on the individual subject headings instead.

In the case of Imaginary creatures, it's LCAC not LCSH, so will not control.

Recently I've noticed that I am unable to highlight data in the Quick Search, Command Line, or Keyword/Numeric Searches. For example, if I want to delete the data and enter a new search, I have to backspace to delete.

This has happened to a few of us. It happened changing from Window 7 to Windows 10, also when changing laptops, not sure why. If you continue to have this problem, reach out to OCLC Customer Support.

Are there any plans to include more information in BFAS about WC Discovery for WorldShare libraries? There are some linkages to BFAS from the Discovery documentation, but not from BFAS out to Discovery. An example would be information on what bib fields and subfields display in Discovery.

Interesting! We are not aware of this linking to BFAS, it will be investigated. We do not have any plans to incorporate that kind of information but will give it some consideration once we can see the linkages that you see.

Should we add Fictitious characters in 650 or 600? I seem to see a mix of both. 

Current practice is to enter them in field 600 rather than 650, which was the older practice. However, if the heading is still in the authority file established as a 650, go ahead and enter it that way so that we are able to switch it in the future when the tag changes.

The French subject headings list (RVM) is still using 650s for fictional characters, I haven't seen a transition to 600s yet.

RVM documentation indicates that they aren't currently planning on switching to 600.

Can you expand on the difference in controlling headings in the different applications (Connexion vs. Record Manager) that Jay mentioned?

That was in relation to the fact that there are additional authority files that will be available in Record Manager that will not be available in Connexion, including Getty Art and Architecture Thesaurus, MeSH and possibly others, these will be available in Record manager, not Connexion.

Since we switched to Windows 10, I have been having problems with macro boxes loading really slowly or macros hanging completely. I did report this to OCLC and tried several things they suggested, but the problem was never resolved. Is anyone else experiencing this? 

It is hard to say, it may be because of how your computer is set up or what version of Windows you are using. It may act differently for different people in different places.. OCLC Customer Support should be contacted.

If you are interested in learning more about macros there are a couple of sights out there that go into Connexion macros, one is Hahn library and

You can make changes to some PCC records and some you can't, why is that?

All of that is outlined in detail in BFAS Chapter 5.2, Member capabilities.

If a detail is changed in a PCC record that doesn't meet the permission level, does it get automatically reverted later?

If it does not meet the permission level you will get an error message that you are unable to make the change, at this time you can report to to make the change for you. Keep in mind, the change may require proof.

We have full cataloging capability. Documentation indicates we should be able to add and edit 505 fields in PCC records. However, there have been several times we have not been able to add or edit 505 fields. Do you know why that would be?

If you are making other changes to the record at the same time and some of those changes are allowed according to your authorization level, but some of those changes are not allowed, you will not be able to replace the record at all.

If you are only making changes to the 505 field, save the record to the online save file and send us a message. We will see if figure out what is going on with it.

It may also be the situation that if you click into a field that are not allowed to edit then click out of it, could possibly recognize it as an edit even though you did not edit it. Make sure you have a fresh copy of the record when you edit the 505 field.

Last month Collection Manager reports were missing due to a glitch. I noticed this morning I did not get any again even though I had files to download.  I tried to log in to Statistics and was unable to do so. Do you know if there is a problem?

I don’t think any of us does know. OCLC does have a status page and that will show all the different systems where there may be maintenance going on. Using the Connexion help menu: HELP/Useful Web Links/OCLC System Alerts, or OCLC Customer support may be better able to answer that question.

Records in your local and online save file, how long do they stay in there? We have some record in that since 2014.

Local save file saves up to a maximum of a large number of records, records in both the local and the online save file will be there until you delete them.

Do you know off the top of your head a very hard work to catalog?

Many interactions of Bible, board games from the '60s & '70s, An 18th-century breviary with multiple pagings and a subtitle containing the names of a whole bunch of popes, any work in a script you're unfamiliar with. Hebrew, Aramaic, Korean, etc., any items that combine lots of smaller works by different authors with only marginal bearing on a particular subject (congresses, collaborative art projects, DVDs).

September 2019: WorldCat knowledge base quality: Connections between the KB and bib records

Topic-specific questions

What is LHR?

Local holdings records (LHR) are attached to a bibliographic record. These types of records allow libraries to provide additional information regarding your item-specific holding. An example of this can be a serial where you can indicate what issues and coverage you have for the title which can be different from the entire run of that serial. Your library may have volumes 5 to 100 even though the serial contains much more.  

How does the format in the bibliographic record (i.e., "Journal," "E-book," "Audio") correspond with the choice of Format in the knowledge base (KB) record? Is it not a 1:1 relationship?

There are some correlations between the bibliographic record and KB record. The type and bib level for a MARC record is described differently and in a more granular way. KBART data is simple and has fewer values to describe resources.  

How can a mismatched link be removed from a KB record? Generally, the title matches but not the author or date. This tends to happen when there is no ISBN for older books.

You always want to include as much information as is available for the resource. For older books it helps to have title, author, publisher, and date of publication. These fields are important for the matching process. There are times though when titles are so similar that erroneous links are generated. If this is the case, you will have to create a separate collection to separate the link from the WorldCat record.

How is the choice of MARC record determined for titles in KB collections?

Generally, the fullest and most correct record is chosen for the resource.

Our institution created the collection: National Academies Press open access monographs Collection ID: global.112903.ScpNAP. When we want to add newly cataloged titles with their URL, etc. info to this collection, is there any way to upload this information in batch, rather than adding titles one-by-one? We have tried to do this but cannot get it to work.

A record load is very different from a collection load. Usually you would need a Batchload service to create new records. Whereas the knowledge base record will show there is a link, set a holding, and so forth as described in the presentation. It seems for this collection you may want to do both.

Also, suggested by an attendee, you could create a KBART file, where you add the OCN in the appropriate field, and then upload to your Collection. Mark confirmed this is possible using the autofill feature but if doing this make sure to also include the title URL in your KBART upload.

Is there any way to have holdings set on records without having an OCLC record number in KB?

There is not.

Within a KB collection, does OCLC run dedupe if different records were added but actually represent the same manifestation?

No, due to the design of the KB database. An example of this is serials where it may seem like duplicate records, but they are not. If there is a true duplicate where everything really does match, then we try and remove one of the records.

Does manually removing an incorrectly matched OCN from a KB title prevent it from re-matching in the future?

If there is an underlying data issue behind that (e.g., such as an incorrect ISBN on the KB record or WorldCat bibliographic record) those will return. We have to do extensive investigative work to see where the incorrect data element is coming from.

If you activate a collection in the KB, does that automatically add your library's holdings (so that it displays in Connexion as "Held by XXX"?

If your settings have been marked in WorldShare Collection Manager to maintain WorldCat holdings then it will.  

Is there an easy way to search Discovery for physical items with an institution's holdings, but no LHR?  I would like to clean up my library's holdings, but the searches I try always result in about 99% eBooks, even when I specify and/or exclude a format type in the search. I have tried excluding eBooks in the search, but they still show up in the search results.

The WorldCat Discovery search criteria that should return library holdings but exclude LHRs and digital content: li:<OCLC Symbol> NOT l4:<OCLC Symbol> NOT x4:digital (Do not include brackets around OCLC symbol). Also, one of the attendees mentioned during the end of August, Discovery was having some search issues--including not matching format limits--which have since been resolved.

Do you know whether there is any plan to create some sort of notification when a new file is being added to the KB? I've had a situation where I saw a certain number of titles in the collection, but then a little while later, there were more titles in the collection. It turned out that the collection was being updated on the back end.

There currently isn’t a notification system besides what is in our activity history. We can talk to our product team to see if this is something that can be considered in the future. Thank you for the recommendation.

Is there a way to get a file of all the records in a collection without waiting for an add or change to any of the records-If the setting is to receive a full set each time?

In WorldShare Collection Manager you can change a setting in the collection to do a daily run and get the records within 24 hours. This would be the best way to do this.

If I change a collection to daily will I get a file in the next day or so?


We've been using the monthly archive of GPO's New Electronic Titles to obtain the OCLC numbers for entering as batch search keys, then we set holdings and export the bibs. I assume that we can do this using Collection Manager. How does the Collection Manager routine work, at least basically? And how would we transition from the monthly GPO file to the titles added since then? For example, the collection U. S. Government Documents at the moment has 161,224 titles.

Yes, this may be achieved by using the WorldCat knowledge base within WorldShare Collection Manager. Autoload feeds are available for libraries that participate in the Federal Depository Library Program (FDLP) where U.S. Government Document collections will be regularly maintained for them. When titles are enabled in this process the library may set up MARC record delivery/output within those collections. You may find more information about how to sign-up for the GPO autoload feed here.

How does merging of bibs affect the KB?

There is a process that accounts for merging. The retained OCLC number will eventually be updated in the KB when merges occur.

If we get MARC records from a vendor, but know the records are in OCLC with good OCLC numbers, is there a way to get OCLC records if the vendor doesn't work with Collection Manager? 

There are several choices to obtain this. You probably want a batch process. You could put together a file from the vendor (if your licensing allows it) and submit them through data sync and get back OCLC records as a result. You can also put together a batch search in Connexion with the unique identifiers (e.g. ISBNs) and export the records to your catalog or just update your holdings if that’s all you need to do. Finally, you can talk to your vendor and let them know it would be great for them to work with us on the knowledge base.  

How long after merging does the KB get updated?

Merged OCNs are applied to the WorldCat knowledge base every 24-48 hours.

When you answer questions that you have to look up (e.g. merging question) where is your answer?

Those answers will appear in the summary and member questions transcript, which can be found at The recording and presentation slides will also be found there.

The National Academies Press (NAP), is a global collection set, is anyone else adding to it?

Any library may add/update/delete titles to any cooperatively submitted collections via the Cooperative Management feature within WorldShare Collection Manager. All cooperative updates go through an approval process before they are loaded into the KB. In cases where the member institutions who are looking to holistically update a global collection they previously contributed, the member library may send a KBART file with their desired collection updates to OCLC Support which can be refreshed by knowledge base team on their behalf. When supplying a KBART file for this process, please include all titles for the collection within the file.

If you have a vendor with its own KB, does this duplicate work done with that vendor's work?

This depends on the vendor services they desire to use. If they want to use multiple services from separate KB vendors then it would be recommended to synchronize the KBs. You may find more information about the 3rd Party mapping process here which enables libraries to import and map their KB holdings from other KB vendors to the WorldCat knowledge base. 

Where is Autofill function located in the KB?

Is an action within the KBART file. It’s the 26th field in the KBART file. You would change it from “Select” or “Raw” (which would mean read and write) to “Autofill.”

October 2019: 1, 2, 3, 4, 5: Everything you wanted to know about stats but were afraid to ask

Topic-specific questions

October 9, 2019

10% merge rate seems low for DDR?

Actually, this is a high rate. When DDR first started, it merged 2 to 3 percent of records it reviewed. This higher rate can be attributed to improvements in DDR processing and the way we ingest batch records now means DDR bears more of the burden to make sure records match as opposed to the records being matched on their way into WorldCat.

In WMS, can I have statistics for a specific cataloger?

Not at this time. This is a request we have received before and we do plan on adding in future developments of cataloging reports.

We don't always re-import the bib when we make BIBCO or CONSER replaces, so we can't use our ILS to track that replace-type stat. We manually keep track of these replace transactions (hash marks!) Does anybody have a more automated solution? 

It sounds like that's a function of what you're doing with your local system. The examples given in chat were from a billing report rather than from the statistics portal, so that would be a different way of tracking which was not covered by today's presentation and outside the scope of what we are able to address.

There is a limit to the number of lines in the stats-reports: will that be lifted any time soon?

With the move to WorldShare Analytics reports the limit to number of lines will be 50,000.  

PCC has asked us to report stats on changed/new series authority records and name authority records separately, which I don't think OCLC Stats currently separate out. Has there been any consideration to separating out these NACO changes?

A further breakdown of the categories has been discussed before. We don't have that in place at this point, but it is something we have been thinking about. If we have the opportunity to revise these statistics in the future it will be considered. 

Could you say more about the different statistics categories available under "WorldCat Record Actions"? For example, what is the difference between Update Existing WorldCat Records, Update New WorldCat Records, and WorldCat Replaces?

Update is about updating holdings, not about updating the record themselves. Update existing WorldCat records would be adding your holdings to existing records in WorldCat. Update new WorldCat records would mean adding a new record to WorldCat and adding your holdings. WorldCat replaces would be using the actual replace transaction to change the WorldCat record.

Are there plans to add new statistics to record manager?

Not at this time. We will be working with the Analytics team in the future to move cataloging reports into WorldShare Analytics but this is not yet scheduled. We will look at improving existing reports in the move and any feedback on cataloging reporting needs is greatly appreciated. 

In WorldShare, I only have access to Metadata and Admin. How do I get access to Analytics? 

Libraries can contact OCLC Support to adjust the roles assigned to librarian’s WorldShare accounts to confirm access to reports is provided.

In WorldShare, under Analytics - Reports, why is my only option E-Resources? 

Libraries can contact OCLC Support to adjust the roles assigned to librarian’s WorldShare accounts to confirm access to reports is provided.

Why are there two categories for Produce (Produce Existing WorldCat Records, Produce New WorldCat Records)? I thought Produce was related to getting cards printed and that that product was no longer available. If that's not what Produce means, what does it mean? If that is what it means, why are those categories still there?

It does relate to producing catalog cards. It is a holdover from the past, we produced catalog cards up through 2015. It sometimes takes several years to deal with the administrative part of changing this system. It's one thing to stop sending out catalog cards, it's another to change these categories to get rid of 'produce'. Basically, it means the same thing as an update, so, produce/update it's the same functionality, it sets holdings in WorldCat for your institution. When you download statistics if you want to combine those categories for your own purposes, you certainly can. The reason for produce existing and produce new, we used to have different billing categories for those back when we had transaction pricing. Now with subscription pricing covering all transactions, there is no longer any distinction other than one is for an existing record and one is for a new record.

Librarian's Toolbox link to OCLC Usage Statistics. The entry for Cataloging covers the topic Using OCLC Usage Statistics for Cataloging which details cataloging reports and cataloging report field definitions. 

How are replaced records counted (# of records or by individual replacement actions)?

Individual replacement actions, so whenever you invoke the replace command.

October 23, 2019

Does this work with Google docs? We no longer have access to MS Excel.

As long as you are able to open or import Excel files into Google Docs, this should work. On the OCLC Usage Statistics page, you can just highlight and copy the information and then paste it into whatever application you are using.

I tried to open the slides from the Oct. 9 presentation but was unable to. What format are they in?

The AskQC presentations are PowerPoint slides. An alternative to Microsoft Office or Google Docs might be to try Libre Office. You may need to speak with someone at your institution who sets up your office applications.

Is there an overlap of the type of reports available on WorldShare and Connexion?

There are very similar statistics but they are not kept in both places. The WorldShare statistics are only available in WorldShare, the Connexion statistics are only available through the OCLC Usage Statistics webpage. While the reports are not exactly the same, they are very similar.

We're a merge library (and we love it a lot!) but our PCC stats don't get recorded on our merge authorizations. Is there a workaround other than doing work in 2 different authorizations (like doing the merge in one and then the authentication in another?)

Yes, you would need to use two different authorizations for that.

Does Record Manager report for a separate set of accounts, or can one get the same report for an account from either Connexion or Record Manager?

The Record Manager reports as shown in the presentation under the Analytics tab reports all the transactions done using the Record Manager interface. It does not include transactions done using the Connexion client or browser interfaces. Another way to say it would be if one person worked in both Record Manager and Connexion and wanted to get a total of their work for the month, they would need to go to both places and combine the numbers.

If you have not used Record Manager, the Connexion authorization numbers do not work in Record Manager. It is an entirely different method of signing on and requires a different set of authorizations.

I listened to the 9 October AskQC session. To confirm .. the update and produce function are basically the same for counting purposes?


In Connexion statistics, when you want to receive reports through e-mail, and you want it sent to more than one address, do you add a semicolon between addresses in the dialog box? Or does each address need to be added separately?

Multiple addresses can be entered in the E-mail Address dialog box. It appears each address can be separated by a semi-colon, comma, or space.

We are a merge library, and I am wondering if it is possible to obtain the merge statistics for our institution? (Unless you already covered this and I missed it?) Thank you.

Yes. if you are a Member Merge library and would like to receive your merge statistics, contact your reviewer or and we will be glad to get you set up with that.

I have a question regarding export. This morning, I tried to submit NACO statistics for midyear report. I got the statistical data. How does the export mechanism work? Someone told me to enter them manually and was given a weblink. But I could not login as my login info was invalid. Whom shall I email?  Where can I get more info about NACO participants submitting statistics?

You will need to take that question to your contact or the NACO contact at the Library of Congress because they run the statistical reporting mechanism for where you enter those statistics. This is totally separate from OCLC and we don't have any control over that. You can get statistics from OCLC, but as for manually entering them, that is a Library of Congress system where you are reporting them.

The general NACO email address at the Library of Congress is

Also, this has been an active topic on the PCC participants list, that is another place you can go for more information.

For NACO questions you can contact Luanne Goodson at

Do we need to contact NACO to set up our stats with OCLC?

No. The earlier question was about reporting NACO statistics to LC. You can get your statistics for what you do using OCLC interfaces, as described in the presentation today.

Does OCLC track instances of controlling headings?

Sort of. We don't keep statistics on controlling headings. We do track headings that get changed which have already been controlled and then if the authority record changes, we get notices of what has changed through the automated updating process.

General questions

October 9, 2019

What automated merge projects have been done on existing records in the last couple of years. (not just recently added records) Perhaps ebooks and eVideos, looking years in different fields, not just in the header. or some other creative area the staff doing manual mergers have seen repeatedly? 

Charlene Morrison outlined an e-serials duplicate project which she and Hayley Moreno presented for the ALCTS Electronic Resource Interest Group at the ALA Annual 2019 ERIG Meeting. They were able to initially identify about 713,000 OCLCE candidate records and merged 157,000 duplicate sets in the first pass. The presentation Multiplicity amending e serials duplication in WorldCat can be accessed here on the ALA Connect website.

We also have a few projects we run on a regular basis with the approval of vendors where we run records through a macro to match on ISBN along with some additional checks. This is usually done on new pre-pub records that normally DDR would not be able to match due to inaccurate pre-publication information.

We have also identified sets of records that we can manually force through DDR when we see patterns we think DDR can now merge.

Is a new record checked for duplication upon creation, or as part of a scheduled scan?

All records that are added to the database online and through batch load are subject to DDR processing as are records that are replaced online. There is a window of about a week after a record has been created or added before DDR processes it. This allows for changes and updates to the record which may happen right away.  

Our library uses Brodart's FLEX account for shelf ready books and MARC records. When they send our holdings to OCLC, sometimes a brief duplicate record is created and our holdings are set on it, instead of being set on a fuller record that already exists for the title. Why does this happen?

These records come in through our DataSync service. If your holdings were set on a brief record as opposed to a fuller record that might already be present in WorldCat, that means the record sent from Brodart didn't match to the existing record in WorldCat for whatever reason and sometimes the reasons are obscure. The paging may be off a little bit, the date of publication may be off a little bit, if there are enough of those little differences it might not match right away. If you believe this has happened, report that to and eventually, the records will be manually reviewed and merged if appropriate. 

Would it be helpful to email the Connection help contact the OCLC numbers to records that are duplicates?

Please send any merge requests to so Metadata Quality staff can merge the records. please see OCLC's Bibliographic Formats and Standards, Chapter 5.5 Requesting Changes to Records for information about reporting change and merge requests to us. There is a section at the bottom of the page that details the different reporting methods available which you may find helpful. 

One new headache we are currently facing is where either ISBN 10 or ISBN 13 has the subfield q but the other does not. Is there a process OCLC can run to create the subfield q when it is missing in one?

Currently, if a record is created online with a field 020 with an ISBN and subfield $q, the subfield $q will carry over when the associated 020 field is created. This appears to now only be an issue with DataSync not retaining the subfield $q along with the rest of the 020 field when transferring the ISBN to records. This is something that would need to be addressed with a DataSync enhancement. It is unknown at this time if the retrospective addition of missing qualifiers could be added to records. 

As the NACO coordinator at my institution, I wish there were a way to view all of an institution's authority records for a time period, e.g., ask Connexion to show me all authority records with our NUC symbol in the 040 from 2018. It would be a game-changer for NACO quality control. You can already search for your symbol as "Cataloging Source" in Connexion, but it shows you all NARs you've done since the beginning of time, which on our case gives 40,000 records and makes my head blow up. Maybe there is a clumsy way to combine Cataloging Source and ARN wildcard search?

Even if we could limit to looking for an 040 $a it would be great and better than what we currently have.

How would you select on last edit date in 005?

The difficulty with this one is the timeframe as the replaced date isn't tied to your symbol in the 040. With our current indexing, there isn't an easy way to do this type of search. A member has suggested limiting your search by the first four digits in the LCCN, such as ln:2018* but that might still give a big group of records.  Authority records are not indexed the same as bibliographic records; however, just about everything is indexed in the keyword index in authorities. So you can do some interesting searching in authorities and retrieve records based on words that appear in something like field 670, which you would not really think of as a traditional access point at all for an authority record.

It does not appear the replaced date is indexed in the keyword index. However, the LCCN is indexed in the keyword search, I was able to narrow searches down to the day with this search kw:20150618* as an example.

There was also the suggestion to replace the dlc in these URLs to your institution's MARC Organization Code: and and search them with a browser. These are RSS feeds, so if you want more eye-readable results, you will want to use an RSS feed application. The output is displayed chronologically.

I'm finding records for non-youth books with one Childrens heading (650 1) when all others are LCSH. Sometimes the Childrens heading dups an LCSH and sometimes it's unique. Many are PCC. Any idea why this happens?

Likely through field transfer, either by the merging of records or via DataSync. When a record comes in and matches an existing record and a particular scheme of subject headings is not present in the record that is already in WorldCat, but is present in the incoming record or the record being merged, that subject heading transfers.

Was there much about metadata at the OCLC Americas Council Library Futures Conference in Phoenix? Any recordings, etc., posted yet?

None of us were there, so we don't have first-hand knowledge about the content presented or of any recordings which may be posted.

Why doesn't the data ingest program run records through the validation process? It's annoying when you make no changes to a record but it won't validate until you correct some field's data.

It is the case that incoming records through data ingest are subject to validation. What we do internally is assign an error level to the results of that validation process, some errors are considered far worse than others. Some may affect the structure of the record itself, which is a problem, so we would not necessarily add those kinds of things. But if errors were more superficial, we would go ahead and add that record to WorldCat figuring it is far better to have the library's holdings represented by a bibliographic record even if there is a problem with a tag or subfield code, than to have no record at all. Of course, we try to fix as many of those as we can, but it is easier to fix records that have patterns of errors than records with an off-the-cuff singleton type of error. This is why you will see records in WorldCat which were added through Batchload will not necessarily pass online validation in the same way that records which were entered directly online will pass validation. We are taking data from another library's system and accept some of the errors that are there. Of course, our goal is to fix them up as much as we can, but we also need your help in doing that.

Has the process that moves the order of subject headings around so they are no longer in order as assigned by the cataloger been stopped?

It is believed this is going to require fixes in multiple locations and this is currently in process.

Might certain pre-pub vendor records be targeted for a merger process, with easier rules for match (i.e. ISBN, fuzzy year, first word or two of title), DDR with specific vendors e.g. Brodart

We are making some enhancements to matching, to match on OCLC control number with some checks in place which should help matching these types of records.

I have started seeing some records with every year the book was published in the 264 4. I had a third edition with 2019, 2010, 2002. Is this going to be a new practice??

Hopefully not.  Please send those to so we can investigate.

October 23, 2019

How do you get to be a library that can merge records?

If you would like to participate in the Member Merge Program, email Laura Ramsey will respond with more details.

Currently, OCLC is harvesting our collections in our DSpace repository. Now we have created a new collection. How can I get this harvested too?

While a great question, this is out of our area knowledge and expertise. The best thing to do would be to contact OCLC Support.

November 2019: Comparisons of searching in Connexion and Record Manager

Topic-specific questions

Are we able to narrow our search to local holdings in Connexion client?

No, local holdings records can only be searched in Record Manager.

Why is there no advanced search box for authority records like there is for bib records in Record Manager? It would be very useful.

This feature has not been implemented yet. The usage of the "advanced search panel" is monitored for bibliographic records. It is believed that this is by far the least used option, after "basic" and "expert" command line searching. So, adding one for authority records, hasn't been a priority, but if we hear more support for it, we will look into it.

Will spellcheck be added to Record Manager?

There are currently no plans to add spellcheck to Record Manager. Instead, relying on the spellcheck that is provided within your browser. The advantage allows us to not have to do spellcheck dictionaries in multiple languages.

What is a Music Hire record?

This is a checkbox found in the bottom right of the advanced search panel in Record Manager. It was added, along with other options, for the work that we did for Te Puna (National Library of New Zealand). At the time, we couldn't think of a way to just show those to users from the National Library of New Zealand. So, we have it there available to everybody but only applies to a few users.

In Record Manager, you can limit book searches to either Print or Online. Why can you only limit Journals searches to Online? I need a limit to Print Journals.

This is a known bug. We will be addressing it with our development teams.

In the example that showed search results in OCLC, the list had columns for number of holdings and control number. My OCLC display does not include these columns. How can I add these columns to my display?

In Connexion client, right-click in the search results list and select List Settings. You can choose which columns you want to display, along with the order you want them to display in by moving them up and down within the list.

Which are the Record Manager roles related to PCC work? Do we need to make new authorization requests or anything like that?

Information on how this works is found in the August 2019 Release Notes for Record Manager.

Is there a way to limit your results in Record Manager to only records that have 856 fields?

In both Record Manager and Connexion, an advanced search you can use "internet resource" as well as the index mt:url. However, this will result in both print and electronic records that contain an 856 field.

Is Boolean nesting available in Expert searches in Record Manager? If not, how are multiple Boolean operators handled?

Yes, you need to make sure that the parentheses are in the correct place within the search. For example: (kw=hrvatski AND ln=rus AND vp:cyr) OR (kw=serbia AND ln=scc)

In Connexion, I use mn: (music number) searches quite a bit. Is there a comparable search for Record Manager?

Yes, this is also available in Record Manager within the advanced search drop downs or expert search syntax. 

Any difference between Connexion and Record Manager on searching non-Latin scripts with same or different results?

Not that we are aware of. A way to test is to use the vp: index label for defined Vernacular scripts. If you find there are differences, please report them.

If you want to search just edition statements in Connexion Client is there a way to do this?

There is no index label for edition statements. Keyword searching using quotation marks is a way to help narrow your search down.

Are there plans for sun-setting Connexion at this point?

No, there is no end-of-life date for Connexion at this point. We continue to study ways to keep it robust in light of evolving releases of Windows operating system. When there is an end-of-life date, it will be announced with ample lead time.

After a search, I would like to download those bibliographic records into one local file. Is it the same in Record Manger, as in Connexion, just to create a new file path?

There are no local save files in Record Manager. We do exports, but no local save files. Depending on the use case, it may be that exporting records to a local file will offer some of the same functionality. If you export it as a MARC record download, it will increment the files in your download folder (e.g., xxx-1.mrc, xxx-2.mrc, etc.). Local processing of those files in MarcEdit is a strong option, as we have integration of the WorldCat metadata API in that product.

Is there a way in Client to find all the eBook records on which we have holdings, or do I have to use Record Manager?

Searching "mt:uri and li:[institution symbol]" or "mt:elc and li:[institution symbol]" will show these results. Depending on the size of your institution's eBook collection, a slash ( / ) and a date, or range of dates, may need to be added to the search.

Can you search the 65x fields for $2 codes in the Connexion Client?

Yes, using the "ho:" index. Indexing information can also be found at the bottom of each field page in Bibliographic Formats and Standards (BFAS).

Is there a way to export records from Record Manager to an ILS such as Alma?

We do offer export into ILS's via TCP/IP export, which is configurable within Record Manager preferences for working with bibliographic records.

In Collection Manager, how can I get the function for "my comparisons" and "managing comparisons" under Collection evaluation?

These are not Collection Manager functions, but instead options under WorldShare Analytics. For more information, please see

I thought some time ago we were able to search in Connexion client authority records for prolific composers and titles and get right to the title.

This should be possible via the Root and Expanded browse functionality. You type in the name of the composer, and there is an option for an expanded term where you would enter the title. This should take you to that title or to the general range where that title would fall if that particular title doesn't exist.

In a follow-up message, one of the attendees at today's VAOH presentation offered the following advice:

When searching in Connexion Client, there are only 3 boxes in the Keyword/Numeric search area. That’s true, but another term can be added to an already-filled-out left-hand box as long as the additional term is preceded by its index label. For instance, let’s assume you need to use all three boxes to get a small enough result set but you also want to limit your result to RDA records.  You can add “dx:rda” to any left-hand box where the index in the corresponding right-hand box is for another index. For example, the following will search for "william" and "shakespeare" in the personal name index, and for "rda" in the 040 $e:

Left-hand box                                                                                                                        Right-hand box

william shakespeare   dx:rda                                                                                                 Personal name (pn:)

About authority browsing in Cnx Client. The advice, which was also mentioned in the presentation, was that after opening the Browse Authority File dialog box, entering whatever you are searching for in the 'Browse for' box, and picking your index, THEN put a word from the part of the index you want to land on in the Expanded Term box.

I tried to do this when I first started using Connexion 15 years ago but gave up after a while because my first search term usually was not specific enough. For example, if I browse for "king, stephen" as a personal name and put "shining" in the Expanded Term box, my result set begins with King, Stephen, 1818-1852. There are two more Stephen Kings before I get to the one I want, which is King, Stephen, 1947- I failed to put the birth date in the search. You have to start over and this time add the birth date.

I've found it easier to enter what you think the name is and just search that. The results make it easy to find which Stephen King you want (the one with more than just one hit!), click once to choose it, enter the Expanded Term at the top, and press Do Expanded Scan or Enter.