richardstartin edited a comment on pull request #7402: URL: https://github.com/apache/pinot/pull/7402#issuecomment-914695077
The ordering makes some difference but the point was we're not entirely in control, so shouldn't over-optimise. I created a worst case scenario, where there's one string, preventing the first branch from being removed and then only `Byte` and other types which fall through to `OBJECT`. It makes a bit of a difference but it's not as bad as the `HashMap`, and is still better than the `ClassValue`. The best case scenario is just strings. So even in the worst case, this does not regress. ``` Benchmark (_scenario) Mode Cnt Score Error Units PinotDataTypeMappingBenchmark.singleValueType worst thrpt 5 280.270 ± 1.529 ops/us PinotDataTypeMappingBenchmark.singleValueType best thrpt 5 309.094 ± 2.198 ops/us ``` (Here are the results with the `HashMap` where performance is better when only one key is ever accessed, I didn't bother to test with `ClassValue` since I just want to establish that this doesn't regress) ``` Benchmark (_scenario) Mode Cnt Score Error Units PinotDataTypeMappingBenchmark.singleValueType worst thrpt 5 167.794 ± 1.517 ops/us PinotDataTypeMappingBenchmark.singleValueType best thrpt 5 204.106 ± 0.434 ops/us ``` -- 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: commits-unsubscr...@pinot.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org For additional commands, e-mail: commits-h...@pinot.apache.org
