Skip to content

Fix treament of dual as sharded by reverting #5268#5519

Closed
aquarapid wants to merge 2 commits intovitessio:masterfrom
planetscale:jacques_revert5268
Closed

Fix treament of dual as sharded by reverting #5268#5519
aquarapid wants to merge 2 commits intovitessio:masterfrom
planetscale:jacques_revert5268

Conversation

@aquarapid
Copy link
Contributor

Fixes #5508

I think we should fix the problem in #5268 by just implementing "SELECT 1 from dual" in vtgate.

@aquarapid aquarapid requested a review from sougou as a code owner December 6, 2019 02:00
This reverts commit 27b5aae.

Signed-off-by: Jacques Grove <aquarapid@gmail.com>
@systay
Copy link
Collaborator

systay commented Dec 6, 2019

+1 #5268 is "just" a performance optimisation, but it's broken quite a few scenarios. I'm currently working on shortcutting dual calls, so this will work soon enough. Until then, this revert is what we should do, IMHO.

/cc @mpawliszyn

@sougou
Copy link
Contributor

sougou commented Dec 13, 2019

Closing in favor of #5551

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Vitess seems to treat inserts with subselects involving DUAL as sharded, even when keyspace is unsharded?

3 participants