Thanks!  That opens things up.  We do have a lot of OCLC numbers.  For my 
example book, there's an 035 with three of them, including 841051199. If I look 
at

http://worldcat.org/oclc/841051199

it takes me to the human-readable page, but

http://worldcat.org/oclc/841051199.rdf

shows it all in RDF, and I can see a lot of things like

<rdf:Description 
rdf:about="http://experiment.worldcat.org/entity/work/data/1613596711#Place/japan">

so I can pick out the work ID and look it up.  (Perhaps the work ID be specified 
directly there?)

So that would work, but aha, I just noticed I could make it a little simpler by 
using xOCLCNUM to get the work ID, which is the owi field here:

http://xisbn.worldcat.org/webservices/xid/oclcnum/841051199?method=getMetadata&format=json&fl=*

And then I can go to

http://experiment.worldcat.org/entity/work/data/1613596711.rdf

and get all the workExample links, and use those OCLC numbers.

(Which I'm sure you knew, Roy, but perhaps didn't mention because of the 
rate-limiting, but as far as I know our subscription means I can get an access 
token so I can do some larger queries.)

A first run of something like this would take a while to process everything, but 
I'd store locally what I need to know, and then incremental updates for a 
month's worth of news ebooks wouldn't take long.  Thanks!

Bill

On 9 December 2015, Roy Tennant wrote:

> Do you have an OCLC number in your records? If so, you could call it at
> WorldCat like this:
>
> http://worldcat.org/oclc/XXXXXXX
>
> scrape the structured linked data on the page, looking for the "Example of
> Work" link, then follow it to the Work Record:
>
> http://experiment.worldcat.org/entity/work/data/1613596711
>
> That then will give you all of the OCLC numbers that we consider are part
> of that work (under the "WorkExample" tab).
>
> I know, not an optimal solution even if you have the OCLC number. But it
> could work if you do.
> Roy
>
> On Wed, Dec 9, 2015 at 1:37 PM, William Denton <[log in to unmask]> wrote:
>
>> I'm looking at how to match print (p) and electronic (e) editions of the
>> same book in our collection.  There is no connection between them in our
>> system (VuFind in front of Symphony).
>>
>> For example, two catalogue entries for two versions of COMPOSING JAPANESE
>> MUSICAL MODERNITY, entirely separate:
>>
>> + https://www.library.yorku.ca/find/Record/3238132
>> + https://www.library.yorku.ca/find/Record/3311584
>>
>> I want know they're the same book so I can do more usage and collection
>> analysis.  I've been looking at two ways of doing it with data available
>> right now:
>>
>> 1 a) MARC 020 (ISBN) can list multiple ISBNs. We have e books where the p
>> editions are listed.
>>
>> 1 b) MARC 776 (additional physical form entry) for e books can list a p
>> ISBN or other control number. If we have that edition, great. If not, need
>> to go from e -> p-we-don't-have -> p-we-do-have, which I could do with
>> xISBN.
>>
>> 2) OCLC's xISBN. When it reports other editions of the same work, it can
>> include e versions.
>>
>> There is also:
>>
>> 3) Vendors supplying data.  For example, YBP seems to have all the p and e
>> editions of books tied together.  We could ask.
>>
>> I've been looking around but can't find any discussion about making these
>> connections.  Have any of you done it?  Know of it being done in code I can
>> see? Written it up?
>>
>> Thanks for any pointers,
>>
>> Bill
>> --
>> William Denton ↔  Toronto, Canada ↔  https://www.miskatonic.org/
>

-- 
William Denton ↔  Toronto, Canada ↔  https://www.miskatonic.org/