Allow users to create unique indexes #385
Conversation
|
Fixes #130 |
|
Hey @mtuncer: putting "Fixes" in a comment won't do anything. It needs to be in the pull request's description. I've added it there. |
| @@ -563,7 +563,7 @@ ProcessAlterTableStmt(AlterTableStmt *alterTableStatement, const char *alterTabl | |||
| * supported for distributed tables and errors out if it is not. | |||
| */ | |||
| static void | |||
| ErrorIfUnsupportedIndexStmt(IndexStmt *createIndexStatement) | |||
| ErrorIfUnsupportedIndexStmt(Oid relationId, IndexStmt *createIndexStatement) | |||
There was a problem hiding this comment.
I think we don't have to pass relationId to the function. Index statement has RangeVar *relation field. We can use RangeVarGetRelid() to get the relationID.
There was a problem hiding this comment.
OK. however it requires getting another lock. Which should be okay.
|
@mtuncer I added some comments to this PR. I'll do the testing after discussing the them. |
| continue; | ||
| } | ||
|
|
||
| attributeNumber = get_attnum(relationId, columnName); |
There was a problem hiding this comment.
get_attnum returns InvalidOid for dropped columns. We could detect it here, but, for DDL command propagation we prefer to get the error from the worker nodes as we do it for other error cases. I am noting it here for documentation purposes.
|
@mtuncer this looks good, also tested once more. Let's |
mtuncer
force-pushed
the
feature/unique_index
branch
from
8bd9c3f to
3c4dadc
Compare
Users can now create unique indexes on partition columns for hash and range distributed tables.
mtuncer
force-pushed
the
feature/unique_index
branch
from
3c4dadc to
e74decf
Compare
Allow users to create unique indexes
Users can now create unique indexes on partition columns for hash and range distributed tables.
Fixes #130