andygrove opened a new pull request, #1641:
URL: https://github.com/apache/datafusion-ballista/pull/1641
# Which issue does this PR close?
Closes #.
# Rationale for this change
The TPC-H benchmark output is hard to scan when comparing runs. Times
print as milliseconds with one decimal place, the default of three
iterations slows down ad-hoc runs, and there is no overall total at the
end. Aligning the per-query format and JSON summary durations with the
datafusion-comet benchmark makes it easier to compare results between
the two engines.
# What changes are included in this PR?
- Print per-query and average elapsed times in seconds with three
decimal places (microsecond resolution).
- Print a `Total time` line at the end of the run, covering all
queries and iterations.
- Change the default `--iterations` value from 3 to 1 for both the
`ballista` and `datafusion` benchmark subcommands.
- Store JSON summary `elapsed` values in seconds so the file format
lines up with datafusion-comet's `durations` array.
# Are there any user-facing changes?
Yes:
- The default of `--iterations` changes from 3 to 1. Pass
`--iterations 3` explicitly to keep the previous behavior.
- The console output now uses seconds (`s`) with three decimal places
instead of milliseconds (`ms`) with one decimal place.
- `elapsed` values in the JSON summary are now in seconds rather than
milliseconds.
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]