jaylmiller commented on issue #5230:
URL:
https://github.com/apache/arrow-datafusion/issues/5230#issuecomment-1435713920
I'm going to be posting notes here as I investigate.
Comparison of using no rows (i.e. the version of SortExec that is on main),
the current PR (i.e. v1 and v2 of this issue), and one in which nothing but
sort_batch is modified (i.e. v1 only on this issue).
```
group no-rows
sort-preserve-row-encoding sort-row
----- -------
-------------------------- --------
sort f64 1.04
524.3±155.42µs ? ?/sec 1.00 505.3±194.42µs ? ?/sec 1.04
526.1±143.65µs ? ?/sec
sort i64 1.00
440.9±24.44µs ? ?/sec 1.11 487.9±639.42µs ? ?/sec 1.01
446.9±290.21µs ? ?/sec
sort mixed tuple 1.00
521.9±92.66µs ? ?/sec 2.07 1078.6±105.77µs ? ?/sec 2.31
1206.5±231.68µs ? ?/sec
sort mixed utf8 dictionary tuple 1.00
521.3±87.21µs ? ?/sec 1.46 759.8±168.21µs ? ?/sec 1.32
688.4±64.57µs ? ?/sec
sort utf8 dictionary 1.04
164.3±13.70µs ? ?/sec 1.10 172.8±152.31µs ? ?/sec 1.00
157.6±27.07µs ? ?/sec
sort utf8 dictionary tuple 1.00
597.8±443.41µs ? ?/sec 1.09 649.7±133.73µs ? ?/sec 1.04
623.7±86.46µs ? ?/sec
sort utf8 high cardinality 1.10
1515.9±434.04µs ? ?/sec 1.00 1372.5±179.96µs ? ?/sec 1.04
1428.0±189.69µs ? ?/sec
sort utf8 low cardinality 1.00
840.3±50.04µs ? ?/sec 1.03 865.8±101.09µs ? ?/sec 1.10
921.9±149.46µs ? ?/sec
sort utf8 tuple 1.00
966.6±168.10µs ? ?/sec 1.36 1318.6±133.52µs ? ?/sec 1.48
1433.4±195.09µs ? ?/sec
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]