Merge sqlite

luisiafalves

Hello,

some time ago i had problems with the syncing of zotero, then i started to use separate databases for each document saving the zotero.sqlite archive.

But now i need to merge 3 of these sqlite archives because i merged 3 documents and need only one reference list.

I tried to just export the databases and than import, but zotero didn't recognize the citations in the text.

Is there something that can be done or i'll have to put all the 200 reference again?

Thanks,
Luísia

dstillman

You should be able to sync each library with the Zotero server in order to merge the data with the identifiers used by the documents intact.

luisiafalves

Thanks Dan Stillman.

The syncing was successful, but when i tried to update the citations in the text i got this error

TypeError: state.registry.registry[state.tmp.count_offset_characters] is undefined

What can i do next?

fbennett

Try switching to another style, then back again.

luisiafalves

Thanks:) It seems to be working when the style is not numeric. but, when i try to update using vancouver I have this same error

TypeError: state.registry.registry[state.tmp.count_offset_characters] is undefined

I tried more than one type of numeric and author-date styles.

What can i do now?

luisiafalves

I tried again others numeric styles and it worked with the BMJ, then i tried again with vancouver superscript but it didn't work. The first error is:

Type Error: value is undefined

and then I try to update again and comes the other error:

TypeError: state.registry.registry[state.tmp.count_offset_characters] is undefined

It's something that can be modified at the style or deleted from the archive?

fbennett

Thanks to your feedback, I think I've spotted a bug in the processor that could cause this. I'll make a fresh processor release (version 1.0.303), and when it's been picked up by Zotero you should be able to test with the development XPI to see if it clears the problem.

(It's unlikely that this was caused by merging duplicates; it looks like a simple processor bug.)

luisiafalves

Thanks again.

But what is this "development XPI"?

And how long before the new version is updated?

adamsmith

The update should happen today.
The development (or "branch") xpi is the version of the plugin that contains all the updates that will be in the next (minor) release - i.e. 3.0.4 - we'd have to see when this gets build, but should also be within the next 24-48hs.
http://www.zotero.org/support/dev_builds#zotero_30_branch

dstillman

The latest 3.0 branch dev XPI has Frank's changes now. If you switch to that to see if this is fixed, be sure to switch back to 3.0.4 when it comes out.

adamsmith

Dan - on that topic - when and how often does the branch xpi get updated and is there anywhere to tell easily (e.g. from the version number) up to which github commit it includes?

dstillman

Since switching to AWS we've just been running the build script manually. I'll try to get auto-building set up again soon.

I'll also have the script append part of the Git commit hash to the version number to give some indication of what commits the XPI includes.

waisman

What if the two databases that you want to merge have a large number of common references? Won't you end up with a merged database that has a large number of duplicates. I have two databases, one that I used to write grants and the other a review article. They have about 80% overlap but I would like to merge them into one database. What do you suggest?

Thanks for your help.

David

adamsmith

There is no easy way to do this. Zotero has duplicate detection, but that is manual on a per-item basis - I don't think you'll get around cleaning your duplicates up semi-manually.

dstillman

But note that if the databases started out as the same (Zotero 2.0 or later) database, items from that original database wouldn't be duplicated. You'll get duplicates of any items that were created independently on multiple computers or that were transferred via export at any point.