[BUG REPORT] Zotero Connector not bring articles over from EBSCO CINAHL

erinwatson · June 15, 2021

Hi,

When I've done a search in CINAHL, I click on the yellow folder icon in the browser bar, and then I select the articles I want to send to Zotero. Although it very briefly displays "Saving", a new message reading, "An error occurred while saving this item. See Troubleshooting Translator Issues for more information" then comes up. Why is this happening? (When I hover over the connector button, it shows that the it is using the "Save to Zotero (EBSCOHost)" translator.

I see that there is a known issue with the items in "my folder" but what I've just described is happening to items that have not been placed in the folder. A student at my library reported having this problem as well.

Thanks very much.

Erin

moserli2 · February 20, 2024

I have the same problem...

adamsmith · February 20, 2024

CINAHL via EBSCO works for me. If you're importing a lot, you may be hitting https://www.zotero.org/support/kb/site_access_limits

otherwise, we'd need exact steps to reproduce (what are you searching, what are you clicking)

erinwatson · February 26, 2024

So, I just tried CINAHL and it worked for me. However, I'm still having trouble with Ovid MEDLINE, Ovid Embase and Ovid ERIC: https://forums.zotero.org/discussion/109984/problem-with-ovid-medline-and-chrome-connector-debug-id-d404826409-have-tried-everything#latest

I just tried Ovid PsycInfo and that doesn't work, either.

travis.taniguchi · April 2, 2024

I have a similar problem that comes up if trying to save more than 50 records. 1-50 are fine. 51+ saves no records and just provides the general error occurred message. Are there any workarounds for this?

adamsmith · April 3, 2024

@travis.taniguchi -- which database exactly? There's a good chance you're seeing this https://www.zotero.org/support/kb/site_access_limits (also includes workarounds) but it's also possible something breaks as we go to the 2nd page of results, which we could fix.

travis.taniguchi · April 3, 2024

I'm searching EBSCOhost. Unfortunately there is not a version accessible without logging in. Each page of returns displays ten results. I can expand the first 5 pages (50 results) and download those successfully. But if I expand the sixth page and try to download 60, it fails. If I just download the references on that sixth page (references 51-60) it works just fine.

Im not sure if this is due to throttling. I was able to download several hundred references using this workaround in a couple hours.

adamsmith · April 3, 2024

Almost certainly throttling. I plaid with CINAHL in EBSCO a bit, and you pretty quickly hit a 429 (too many requests) error.
If you submit an error report ID from the Zotero connector after hitting that error we can confirm, but I'd be very surprised if that wasn't the issue.
I think the exact number of 50 you describe is a co-incidence and not what I'm seeing.

Note that the workaround on the site access limit page -- collecting items in an EBSCO folder and then exporting that to RIS or getting a download link for the full search results up to 25k items -- will work well.

travis.taniguchi · April 3, 2024

I conducted a new search and submitted the report.

Report ID: 325131967

adamsmith · April 3, 2024

@dstillman could you take a look?

adomasven · April 3, 2024

[JavaScript Error: "HTTP request to https://research-ebsco-com.evergreen.idm.oclc.org/c/oj3zbi/search/details/z3yv2n6qtj?limiters=None&q=corrections has timed out after 15000ms"]

Nearly a hundred lines like this, which is probably throttling after having received 429, although I don't see it in the report log.

adamsmith · April 3, 2024

Ah, this is a different EBSCO service (we refer to this as the discovery layer, not standard EBSCO) -- I actually don't seem to get 429d there as easily, but also can't replicate the issue: I was just able to get 90 items at once, so again, I would indeed suspect some throttling.

travis.taniguchi · April 3, 2024

Let me know if I can provide any more info that would help troubleshooting. Otherwise I will just use the workaround. Thanks for the help!