Hacker News Reader

Comment by benjamir | Hacker News Reader

We tried to migrate a reference library with more than 17k entries, but it never worked out beyond 2k.

Sorry to say: might be fine for small personal projects, for anything serious I don't see how.

Something's not right there. People use Zotero with libraries with tens of thousands of items all the time. (Some people have hundreds of thousands of items, though that gets dicier.) Not sure exactly what you were seeing, but we'd be happy to debug in the Zotero Forums if you wanted to try again. It's possible to configure a third-party plugin such that it affects performance, for example.

(Disclosure: Zotero dev)

I’ve been using Zotero since it was in beta in 2007, and have a library with tens of thousands of items. As dstillman says that’s hardly unusual. It works perfectly (it’s just a SQLite db with a mature, well-designed schema after all), and syncs perfectly across five devices.

Funny, we tried probably your solution for imports? urschrei/pyzotero is yours? Failed all the time. To be fair we got a lot of 5xx, so maybe we were just out of luck. Anyway time ran out with that project.

I can imagine many reasons why attempting to create 17k items via the web API without some kind of batch processing and/or rate limiting may not have been a successful endeavour.

Isn't zotero backed by sqlite db? So these row counts are nothing much to cope with, sqlite would do just fine.

Do you mean it did not work out for you performance-wise or the export process failed?

What problems did you encounter with this large library?

The import via API failed too often.

What were you trying to migrate from? And what do you use now?

Reply on news.ycombinator.com