Why are records found in WorldCat.org not available in the WorldCat Search API?

The WorldCat Search API includes all the records found in the cataloged WorldCat database.  These are the records found within the Connexion WorldCat database or in the FirstSearch WorldCat database.  It is all the records contributed to OCLC through our cataloging libraries. 

The WorldCat.org database includes all of these cataloged WorldCat records and also many other records, primarily articles, that have been added to this database from other sources.  Those records not in the cataloged WorldCat database display in WorldCat.org with information about their source in the label “Database”.  While they have OCLC record numbers that will retrieve these records in WorldCat.org, they are not part of the WorldCat Search API database. 

Why are keyword search result counts different, in the API, FirstSearch, and WorldCat.org?

There are a couple of reasons for this. WorldCat.org includes article citation and digital materials descriptions that are not searched in the API or FirstSearch interfaces to the WorldCat database, so its results tend to be larger. The keyword index in WorldCat.org currently has a more extensive scope than that in the API and FirstSearch. We're looking into whether the API keyword index should be changed to match the scope of WorldCat.org. The API and FirstSearch results tend to be closer in number, but the API result counts will often be somewhat lower because the results are organized to group items with the same authors and titles together, rather than in individual bibliographic records.

Does the relation operator "<>" work in SRU CQL queries?

It doesn't, and the SRU Explain interface has been updated to reflect this. Not all of the CQL relations are supported. It is still possible to use the "NOT" boolean operator with two separate indexes. E.g., to eliminate results of the material type Books from a search for the keyword "xml": srw.kw = "xml" NOT srw.mt = "bks"

Results sorted by Score appear in ascending order. Could the default be descending order, to put the most relevant matches first?

Ascending order is currently the default for all sorts other than relevancy. It's a sensible default for Author and Title, for example, but isn't for some others. The API currently doesn't apply different default orders for different Sort keys, but that's something we can look into. For now, a sort order can be changed to Descending by adding ",,0" to the Sort key, e.g, "Score,,0".

If a user is from a recognized on-campus IP address, can they be considered as a full service level API user?

Yes they can. Conventional methods of authorizing users for other licensed resources, which include IP authentication, are applicable for determining which users can access the API at the full service level.

What are the limits currently encountered in the service?

  • Number of queries per day
    There is a limit to the number of queries available within a single day set for each wskey. Initially all Member's keys have a limit of 50,000 queries a day set as a rolling 24 hour limit. This can be raised by any member with a key that needs to have more access for a particular use of the API. If your key needs to be increased please send a request to OCLC support staff  via our contact form.
  • Maximum number of records/libraries –
    For any query the maximum number of records or library locations that can be requested is 100. It is possible to send another query with the next start position and request the next set of records or library locations up to another 100.
  • Total maximum number of records for a single specific query –
    For any specific query, it is possible to page through all the results up to the 10,000th record. After that point, the system will appear to continue to page, but the next result will be a repeat of an earlier record.

Is it possible for an institution to have more than one wskey?

Yes, it is possible and even ideal for a variety of reasons. There may be more than one IT group involved in different projects. Or it may be that you’d launch different applications with different keys. If you already have a key and would like another one, please send a request to the OCLC support staff via our contact form.

Why are some libraries that hold an item not appearing as library location information?

The WorldCat Search API provides different levels of holdings depending on the service level of the request. All queries that don't request results at the full service level will return the default level results.

A library location query at the default level provides information only on libraries that participate in WorldCat.org and choose to have their holdings visible. Library location requests sent with the full service level provides information on all libraries' holdings. If your library would qualify to display but does not, it maybe that your library decided at one time to opt out of the general display of their holdings on WorldCat.org. This is a profile setting for your library, which can be changed if so desired.

The service level that should be requested depends on whether or not the application/services' end user is an authenticated library user. The full service level access is limited to direct library users (such as staff, students of an academic library, patrons of a public library, and in-house users). This is similar to the way other licensed databases' limit access to a library's authenticated users and authentication can be determined in the same ways.

More information on the differences between service levels can be found at - http://www.oclc.org/developer/documentation/worldcat-search-api/service-levels.

The different level of holdings information is also seen in WorldCat.org. Any users of WorldCat.org which are known to belong with a qualifying library can see all holdings. But any users that are not known to be affiliated with a qualifying library will see only the holdings information for libraries that participate with WorldCat.org. Many users coming from the library’s IP address or through a proxy server will see the full results in WorldCat.org as an authenticated user. However, coming from another IP address or without the proxy server, sees what would appear in the API as the default level view of holdings.

What does it mean to have FRBR grouping on or off?

In WorldCat there tends to be a different bibliographic record for each new publisher or edition or type of format for a title. Because too many library users there can appear to be very little difference between these, FRBRgrouping is offered to reduce the appearance of duplicates.  FRBRgrouping pulls together bibliographic records with the same authors/creators and titles. Then the group of records is represented by the most widely held bibliographic record for that grouping.  It also provides all the library holdings information for the group of records as one set of holdings.

While this generally works well, there are imperfections in pulling together all these bibliographic records into “FRBRized” groups. The most common reasons are the presence or absence of a uniform title or differences in the authors.  Also realize that FRBR groups often include multiple document types represented by a single record, such as audio-books or ebooks included with printed books.

When selecting FRBRgrouping=off, the results will give each bibliographic record, without trying to remove the perceived duplicate titles for each edition. FRBRgrouping=off is also available for library holdings information, providing holdings information for each bibliographic record only and not indicating libraries that hold a similar edition.