Script: Soft delete unused poste_urls

[molly-moen]

PLC-937
Script to set all poste_urls to have a deleted_at value except those in the list in poste_url_constants.

Links

• jira

Testing story

I verified the script soft deletes all urls in poste_urls except those in the list. I ran the script against a sample table with about 2000 rows. Suresh ran an update against the whole table on a production clone and it took 1 minute 22 seconds.

Reviewer Checklist:

• Tests provide adequate coverage
• Code is well-commented
• New features are translatable or updates will not break translations
• Relevant documentation has been added or updated
• User impact is well-understood and desirable
• Pull Request is labeled appropriately
• Follow-up work items (including potential tech debt) are tracked and linked

[bencodeorg]

Is there opposition to just running:
UPDATE poste_urls SET deleted_at = NOW()
Then doing the update step (maybe do that via the script you have here)?

I do have a vague recollection of our production DB having a setting preventing updates without where clauses, but looks like that might be avoidable?

[molly-moen]

Is there opposition to just running:
UPDATE poste_urls SET deleted_at = NOW()
Then doing the update step (maybe do that via the script you have here)?

I do have a vague recollection of our production DB having a setting preventing updates without where clauses, but looks like that might be avoidable?

@bencodeorg The opposition was the concern trying to update all 3.5 million rows in one shot could have some performance impacts, this way we do one at a time. Also the issue you noted about not being able to update without a where clause.

[sureshc]

Do we have a way to test this? I think Sequel will generate a long WHERE condition (id IN [x,y,z,d,.........]) which might generate a SQL statement that’s excessive in length and which might be inefficient to carry out the multi-row update. We might have to iterate over all of the URLs to that we want to keep and update each one individually.

[molly-moen]

I tested locally (using the same list of ids) and it seemed to work fine, although it was against a smaller table. I can update to iterative if it's safer.

[sureshc]

As long as the list of IDs was the same, we’re good. We’re not writing to this table anymore, so even if the UPDATE of a long list of rows selected by ID is slow, we’re fine.