Add some more aggregate sqllogictests and remove rust tests #4505
Conversation
| # TODO: fix decimal places | ||
| query TIR | ||
| SELECT c1, c2, AVG(c3) FROM aggregate_test_100_by_sql GROUP BY CUBE (c1, c2) ORDER BY c1, c2 | ||
| ---- |
There was a problem hiding this comment.
Do we still need to print so many lines? How about using limit to control the output? cc @mvanschellebeeck @alamb
There was a problem hiding this comment.
Yeah this was just a migrated test - should we track simplifying some of these tests in a separate PR/issue?
There was a problem hiding this comment.
Simplifying during migration or in separate PR both ok to me. -- Seems during migration can reduce our workload
There was a problem hiding this comment.
👍 - simplified in 41fe83f35300
There was a problem hiding this comment.
For large outputs there is some sort of sqllogictest standards:
https://duckdb.org/dev/sqllogictest/result_verification
mode output_hash
And
https://www.sqlite.org/sqllogictest/doc/trunk/about.wiki
The "hash-threshold" record sets a limit on the number of values that can appear in a result set. If the number of values exceeds this, then instead of recording each individual value in the full test script, an MD5 hash of all values is computed in stored. This makes the full test scripts much shorter, but at the cost of obscuring the results. If the hash-threshold is 0, then results are never hashed. A hash-threshold of 10 or 20 is recommended. During debugging, it is advantage to set the hash-threshold to zero so that all results can be seen.
There was a problem hiding this comment.
I reviewed the tests that were removed from rs and verified that there were corresponding tests in the .slt file. Thank you @mvanschellebeeck
| # TODO: fix decimal places | ||
| query TIR | ||
| SELECT c1, c2, AVG(c3) FROM aggregate_test_100_by_sql GROUP BY CUBE (c1, c2) ORDER BY c1, c2 | ||
| ---- |
There was a problem hiding this comment.
For large outputs there is some sort of sqllogictest standards:
https://duckdb.org/dev/sqllogictest/result_verification
mode output_hash
And
https://www.sqlite.org/sqllogictest/doc/trunk/about.wiki
The "hash-threshold" record sets a limit on the number of values that can appear in a result set. If the number of values exceeds this, then instead of recording each individual value in the full test script, an MD5 hash of all values is computed in stored. This makes the full test scripts much shorter, but at the cost of obscuring the results. If the hash-threshold is 0, then results are never hashed. A hash-threshold of 10 or 20 is recommended. During debugging, it is advantage to set the hash-threshold to zero so that all results can be seen.
| @@ -274,187 +271,290 @@ SELECT approx_distinct(c9) AS a, approx_distinct(c9) AS b FROM aggregate_test_10 | |||
| ---- | |||
| 100 100 | |||
|
|
|||
| # TODO: csv_query_approx_percentile_cont | |||
| ## This test executes the APPROX_PERCENTILE_CONT aggregation against the test | |||
There was a problem hiding this comment.
👍 thank you for porting this over
|
Benchmark runs are scheduled for baseline = c806cd1 and contender = 237233f. 237233f is a master commit associated with this PR. Results will be available as each benchmark for each run completes. |
Ports more aggregate tests from aggregate.rs and removes the existing ones.
Work towards closing #4495