collectiveaccess / providence

Cataloguing and data/media management application
GNU General Public License v3.0
290 stars 166 forks source link

ExistingRecordPolicy in Data Import not functioning as expected (develop) #1525

Open mbutel opened 6 months ago

mbutel commented 6 months ago

We're encountering a persistent issue with the existingRecordPolicy in data import, where various policies like merge_on_preferred_labels_with_replace, merge_on_idno_with_replace, and merge_on_idno_and_preferred_labels_with_replace are not functioning as intended. Despite configuring the system to avoid duplicates, these policies fail to update existing records properly.

Several attempts were made, including using different mappings and datasets, ensuring correct declaration of type identifiers for objects (item) and entities (ind), and even changing locale settings, but the problem persists. The issue is observed in version 1.8 of the software.

Responses from other users indicate similar challenges, with attempts to import relationships between objects and entities also failing. Despite following suggested solutions like setting the existing record policy to merge_on_idno and adding ignoreTypeForExistingRecordPolicy, the problem isn't resolved. Debugging-level logging doesn't reveal any errors or warnings.

A potential issue with the mapping of relationship types and identifiers has been noted, where relationship types are incorrectly used for other elements. Some users have resolved related issues by formatting Excel columns from date to text and following contents from previous forum discussions.

The primary concerns are:

The original discussion can be accessed here: https://webcache.googleusercontent.com/search?q=cache:YJJ4uNTUFF8J:https://collectiveaccess.org/support/index.php?p%3D/discussion/301304/existingrecordpolicy-ignored-in-data-import&sca_esv=592607226&hl=en&gl=ar&strip=1&vwsrc=0

collectiveaccess commented 6 months ago

Please provide current, detailed, examples of the problem(s) you are encountering and we'll see if we can reproduce the issue. We will need the mappings and sample data. In the past, we were not able to reproduce the errors you describe.

Please also be specific about the revision of the develop brand you are using.

behindicebergs commented 6 months ago

I've been having the same issue, on version 1.8 I'm unable to run imports regardless of existingRecordPolicy. I've already checked in the config files that duplicates are allowed and the user has root access to MySQL. I've tried with idno and preferred label, and it's always the same, even trying with CVS and XLSX formats. In a couple days I'll upload mappings and source data along with the error message.