Now that we are doing all the DB interactions via the Job Queue, there's really no reason to super-optimize the queries. We should remove the denormalization of the page information in the schema and just use the page_id. This may also require modifying some of the queries.
Description
Description
Details
Details
Subject | Repo | Branch | Lines +/- | |
---|---|---|---|---|
Normalize database for PageAssessments | mediawiki/extensions/PageAssessments | master | +21 -33 |
Status | Subtype | Assigned | Task | ||
---|---|---|---|---|---|
Resolved | None | T120219 PageAssessments deployment to WMF wikis | |||
Resolved | Niharika | T124798 Normalize PageAssessments DB schema |
Event Timeline
Comment Actions
Aaron says that joining with the page table isn't very expensive and it will probably be better to just use the page_id.
He also proposed a couple of other schema changes:
https://gerrit.wikimedia.org/r/#/c/266584/
https://gerrit.wikimedia.org/r/#/c/266583/
Comment Actions
Change 267638 had a related patch set uploaded (by Niharika29):
Normalize database for PageAssessments