adapter: correctly report custom types in mz_columns #27318
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
sploiselle
approved these changes
May 28, 2024
| .. | ||
| } => { | ||
| let entry = self.get_entry(custom_id); | ||
| // NOTE(benesch): the `mz_columns.type text` field is |
There was a problem hiding this comment.
fwiw regtype should resolve the OID to the right entry
There was a problem hiding this comment.
Yeah, and you can join mz_columns.type_oid against mz_types.oid by hand! It's just at odds with the rest of the mz_catalog tables that you have to do the join with OID compatibility layer instead of being able to do it with Materialize's native global IDs.
|
Got a little spooked that no tests failed, so pushed up a new test file with some cases that would have failed before this change. |
morsapaes
reviewed
May 28, 2024
For columns with a type of list, map, or record, the mz_columns table
was incorrectly reporting the generic psuedotype for the column, even
when the type was a named type in the catalog.
For example, given the following objects:
CREATE TYPE my_rec AS (x int, y int);
CREATE TABLE my_tab (f1 my_rec);
the mz_columns table would incorrectly report that column f1 had type
record (OID 2249), rather than type my_rec with a dynamic OID determined
at envd boot.
This commit is a surgical fix that fixes the immediate issue.
In the future, we may want to consider a deeper fix that makes the
conversion between a `ScalarType` and a `pg_repr::Type` less prone to
this sort of bug. That is a much larger refactor, however.
We'll also want to deprecate `mz_columns.type` in the future, for the
reasons noted in the comment in the patch. That too is left to future
work for expediency.
8 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
For columns with a type of list, map, or record, the mz_columns table was incorrectly reporting the generic psuedotype for the column, even when the type was a named type in the catalog.
For example, given the following objects:
the mz_columns table would incorrectly report that column f1 had type record (OID 2249), rather than type my_rec with a dynamic OID determined at envd boot.
This commit is a surgical fix that fixes the immediate issue.
In the future, we may want to consider a deeper fix that makes the conversion between a
ScalarTypeand apg_repr::Typeless prone to this sort of bug. That is a much larger refactor, however.We'll also want to deprecate
mz_columns.typein the future, for the reasons noted in the comment in the patch. That too is left to future work for expediency.Motivation
This PR fixes a previously unreported bug.
This is preventing the dbt-materialize adapter from properly enforcing contracts against custom record, list, and map types.
Checklist
$T ⇔ Proto$Tmapping (possibly in a backwards-incompatible way), then it is tagged with aT-protolabel.mz_columns.typeandmz_columns.type_oidwhen the column's type is a custom named record, list, or map. Previously, Materialize reported the generic psuedotype (record,list, ormap, respectively); now Materialize reports the specific name of the custom named type.