You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Unconfirmed: but it seems that in the case of multi-field queries, some books may be being duplicated; so if you're searching for EITHER "whaling" or "fiction" in subject, say, Moby Dick's word count might be counted once for whaling AND once for fiction, which is obviously extremely bad behavior.
The fix is actually somewhat difficult; probably the way to do it is with a materialized table operating on an "IN" query rather than a join, which:
Is only possible on MySQL 5.6 or later, I believe, with an automatically derived index on the materialized table, which is the only way this will be fast.
Will slow down, probably only a little, any queries involving it since the index needs to be built anew. Depending on the speed effect, it might be worth only applying this to queries that use non-unique metadata items.
The text was updated successfully, but these errors were encountered:
Unconfirmed: but it seems that in the case of multi-field queries, some books may be being duplicated; so if you're searching for EITHER "whaling" or "fiction" in subject, say, Moby Dick's word count might be counted once for whaling AND once for fiction, which is obviously extremely bad behavior.
The fix is actually somewhat difficult; probably the way to do it is with a materialized table operating on an "IN" query rather than a join, which:
The text was updated successfully, but these errors were encountered: