I can answer that question -- they are not multi-threaded.  Search is done via 
Z39.50 or SRU (though, the record itself must be pulled via the API in order to 
get the IDs), and then updates must happen on a record by record bases (so 
again, one call per record) via the API.  Each record is handled individually, 
and procedurally because the ability for alma to handle threaded updates seemed 
to vary by institution.  Taking the slower (but more reliable approach) seemed 
best.

You can see the API code here: https://github.com/reeset/alma_api_library

FYI -- I'll be changing it slightly to accommodate a few changes to the Alma 
api happening in September.

--tr

-----Original Message-----
From: Code for Libraries [mailto:CODE4LIB@LISTS.CLIR.ORG] On Behalf Of Steve 
Meyer
Sent: Monday, August 28, 2017 12:20 PM
To: CODE4LIB@LISTS.CLIR.ORG
Subject: [CODE4LIB] Are MarcEdit Alma API requests multi-threaded?

We are investigating using Marc4Edit to batch update MARC records in Alma via 
the Alma APIs.

I realize this question is likely more appropriate for the MarcEdit listserv, 
but I am asking from the perspective of our systems and development 
departments, not as the MarcEdit users in our cataloging departments. We are 
concerned about API throttling, quotas and the impact on multiple API requests 
against Alma.

The Alma APIs are rate limited; so before embarking on this approach, we'd like 
to make sure we don't hit our API quotas. Can anyone tell us if Marc4Edit's 
Alma bib record write operations are multi-threaded, and if so, is there any 
way to configure the number of threads spawned?

We could end up in a situation where other critical systems that communicate 
with Alma via its APIs could become unavailable due to our API account being 
flooded with what is actually a lower priority set of requests.

Thanks,
Steve

Reply via email to