jackyhu-db opened a new pull request, #3189:
URL: https://github.com/apache/arrow-adbc/pull/3189
## Motivation
When the client calls `HiveServer2Statement` to execute metadata query (e.g.
GetCatalogsAsync, GetSchemasAsync), it won't set `_directResults`, which leads
to a unnecessary thrift call (`CloseOperation`) in its
`HiveServer2Statement.Dispose`, this is because it will make a blocking call
`CloseOperation` if `_directResults=null` (see code
[here](https://github.com/apache/arrow-adbc/blob/main/csharp/src/Drivers/Apache/Hive2/HiveServer2Statement.cs#L378)),
it introduces 100ms-200ms latency for every metadata call/query
## Change
Set `HiveServer2Statement._directResults` from resp.DirectResult in every
metadata API in `HiveServer2Statement`, thus thrift `CloseOperation` won't be
called in `Dispose` if metadata query/option returns a successful
`DirectResult`.
## Testing
- All the E2E testing on Databricks driver
- PGPerf.exe and saw 200-300ms performance improvement for every query
scenario with Databricks
--
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]
