Phil Rzewski

After coming across more queries that trigger the "ambiguous column reference" message, I've got some easier repro recipes. I can't say with certainty that these all share the same root...

Here's a summary of a recent group discussion on this topic. Revisiting the original repro query with current super commit 857796a, @mccanne explained that the error is because when operating...

Here's a new twist I just stumbled upon: It looks like leveraging the static schemas from Parquet files is currently dependent on the presence of `.parquet` extensions. Picking up where...

The changes in linked PR #6341 improve on some examples shown above, as now schema information is obtained from Parquet files even if the extension is not `.parquet` and there's...

Here's a variation in [another sqllogictest](https://github.com/duckdb/duckdb/blob/d524c57a58c4984406ca7ee7251bb8603614eb7a/test/sql/join/left_outer/test_left_outer.test#L84-L90). This uses the same input data as in the first repro above. Due to the join condition `1=2` always evaluating to `FALSE`, in Postgres...

Update: The changes in #6264 have brought the query results here closer to matching SQL, but not all the way. Whereas before the query results from `super` had empty fields...

In a recent [community Slack thread](https://super-db.slack.com/archives/CTSMAK6G7/p1760626403618189) a user reported bumping into this issue.

Here's another simple example based on [another sqllogictest](https://github.com/dolthub/sqllogictest/blob/a1f4aa3c1d691552a4a03d9603c45bcf93d0bb9d/test/random/groupby/slt_good_12.test#L10393-L10395). Repro is with super commit 4493acd. When attempting to negate the result of an aggregation that contains `DISTINCT`, it fails to parse....

@mccanne gave this one a quick look recently and sensed the fix may just need a parser change.

I'm pleased to report that the changes in #6453 have at least addressed that last example in the most recent comment: The `DISTINCT -83` in that query is no longer...