#36632: AlterIndexTogether in historical migration drops overlapping indices
-------------------------------------+-------------------------------------
     Reporter:  Michael Herrmann     |                    Owner:  (none)
         Type:  Bug                  |                   Status:  new
    Component:  Database layer       |                  Version:  5.1
  (models, ORM)                      |
     Severity:  Normal               |               Resolution:
     Keywords:  index_together       |             Triage Stage:  Accepted
    Has patch:  0                    |      Needs documentation:  0
  Needs tests:  0                    |  Patch needs improvement:  0
Easy pickings:  0                    |                    UI/UX:  0
-------------------------------------+-------------------------------------
Changes (by Jacob Walls):

 * keywords:   => index_together
 * stage:  Unreviewed => Accepted
 * summary:  AlterIndexTogether drops overlapping indexes =>
     AlterIndexTogether in historical migration drops overlapping indices
 * version:  5.2 => 5.1

Comment:

 Thanks for the clear report. Bisected to
 2abf417c815c20f41c0868d6f66520b32347106e. There was a follow-up to that
 removal with respect to fixing a crash in
 b44efdfe543c9b9f12690b59777e6b275cb08103, but it looks like we did not go
 as far as fixing what you report here: that the correct indexes are not
 recreated following a table remake.

 It looks like Simon sketched out a possible implementation on the
 [https://forum.djangoproject.com/t/django-5-1-alterindextogether-raising-
 typeerror-class-meta-got-invalid-attribute-s-index-together/36450/3 forum]
 before a slimmer implementation was chosen to fix the TypeError:

 > We can definitely adapt AlterIndexTogether.state_forwards and
 .database_(forwards|backwards) to delegate to AddIndex and RemoveIndex by
 introspecting state and from_state though.

 > Whichever solution we decide to go forward with should serve as a
 validation for attempting the same deprecation path on unique_together ->
 constraints.

 #31834 is almost identical for `unique_together`, but it's possible they
 could be (might have to be?) solved separately, since here we're talking
 about an operation that can only exist in historical migrations.
-- 
Ticket URL: <https://code.djangoproject.com/ticket/36632#comment:1>
Django <https://code.djangoproject.com/>
The Web framework for perfectionists with deadlines.

-- 
You received this message because you are subscribed to the Google Groups 
"Django updates" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion visit 
https://groups.google.com/d/msgid/django-updates/0107019995e16408-aac15a3a-cf56-4e34-83d7-b13a5887904c-000000%40eu-central-1.amazonses.com.

Reply via email to