DEV: Correct historical schema discrepancies (PR #9955)

GitHub

This is very cool!

Overall this is great and should be very safe to merge.

One change though I would like to see is to query the information schema in batches, its a small change, but means this migration in the happy state will only run 3 selects vs 30.

Once that is done merge away.

select * from (
   select 1 id, 2 x_id
   union all 
   select 2, 4
) as X 
join table...

Or a CTE or something.