zed
zed copied to clipboard
brimdata
A novel data lake based on super-structured data
Zed may benefit from having a "strict mode" option that requires programs to be written in canonical form (#2197) in order to be parsed and executed. ## Details Zed's shorthand...
In #4308 review, @mccanne noted: > We should add multiline comments (`/* ... */`) at some point. I think there was a problem a long time ago with the PEG...
While refactoring the optimizer, we realized the code would be simpler if we required all paths of a dag.Parallel to be dag.Sequential instead of dag.Op. Also, parallel implements two semantics...
A user attempted to call operators from within a user-defined `func`, but the parse error produced was generic and did not point at the operator. ## Details Repro is with...
A user asked in a community Slack thread: > Here's sort of a related situation I just ran across. Using the spread operator with a mismatch of array or record:...
A user asked in a [community Slack thread](https://brimdata.slack.com/archives/CTSMAK6G7/p1720194738446819): >I presume this is more or less expected, that if there's a name collision between a type and a field, the type...
This commit removes double slash comments to align with SQL as we already support SQL style comments (--).
This commit enables multi-line comments in the PEG grammar. We had a past issue with making this work but it just works now in the current grammar.
Repro is with super commit 9449896. This was found via a query from [a sqllogictest](https://github.com/duckdb/duckdb/blob/0cbb56d14cc1464d945a06cb8bac051485fa2ef7/test/sql/aggregate/group/test_group_by.test#L151-L156). Repro with test data: ``` $ cat integers.sup {i:3(int32),j:4(int32)} {i:3(int32),j:4(int32)} {i:2(int32),j:4(int32)} $ super -version Version:...