When executing a batch delete request with _creationTimeUnix filter, I am getting errors coming back with message batch delete objects: cannot find objects: finding matching doc ids in shard "xxxxxx": could not get uuid from doc_id=xxxxx
Delete request looks like:
curl -XDELETE -H "content-type:application/json" $weaviateUrl/v1/batch/objects -d '"{"match": {"class":"ClassName", "where":{"operator":"LessThanEqual", "path":["_creationTimeUnix"], "valueText":"1705519625249"}},"dryRun":false}'
It looks like the entire batch fails with even in one ID mapping fails.
Is it possible for the batch code to return back errors and still execute on the successful IDs? What causes the ID mapping to fail? Is there something going wrong at index time or is this strictly a query time issue?
Using weaviate database version 1.23.1
After digging into the source code looks like single ID mapping fails the entire batch in file: weaviate source code
Duda from Weaviate here. This is most certainly a bug. Our team will investigate this.
We will open a bug in https://github.com/weaviate/weaviate/issues
Meanwhile, if deleting a single object instead of in batch, does it work as expected?
Edit: the issue is open
Thanks!