[MINOR] Respect Spark bloom filter config

[jinchengchenghh]

Respect Spark four configs, if not exists, Velox will use its default config value.

spark.sql.optimizer.runtime.bloomFilter.expectedNumItems
spark.sql.optimizer.runtime.bloomFilter.maxNumItems
spark.sql.optimizer.runtime.bloomFilter.numBits
spark.sql.optimizer.runtime.bloomFilter.maxNumBits

[github-actions]

Run Gluten Clickhouse CI on x86

[FelixYBW]

We may remove the gluten config if there are existed ones for Spark.

[github-actions]

Run Gluten Clickhouse CI on x86

[github-actions]

Run Gluten Clickhouse CI on x86

[github-actions]

Run Gluten Clickhouse CI on x86

[github-actions]

Run Gluten Clickhouse CI on x86

[github-actions]

Run Gluten Clickhouse CI on x86

[github-actions]

Run Gluten Clickhouse CI on x86

[FelixYBW]

how about config:

spark.sql.optimizer.runtime.bloomFilter.applicationSideScanSizeThreshold
spark.sql.optimizer.runtime.bloomFilter.creationSideThreshold
spark.sql.optimizer.runtime.bloomFilter.maxNumItems

[github-actions]

Run Gluten Clickhouse CI on x86

[jinchengchenghh]

spark.sql.optimizer.runtime.bloomFilter.applicationSideScanSizeThreshold
spark.sql.optimizer.runtime.bloomFilter.creationSideThreshold
These above two configs decide if insert bloom filter for the join, which take effect in Spark java code.

spark.sql.optimizer.runtime.bloomFilter.maxNumItems
velox does not have corresponding config, I will check if we can respect the config in Velox.

[jinchengchenghh]

After spark3.4 add this PR, the num bits algorithm is changed, apache/spark#37697, and it consider the max num items config, I will update Velox function to respect this behavior

[jinchengchenghh]

Before that PR, max num items is only used in checkInputDataTypes(Spark analysis) to check the config expectedItems is reasonable

[FelixYBW]

spark.sql.optimizer.runtime.bloomFilter.applicationSideScanSizeThreshold spark.sql.optimizer.runtime.bloomFilter.creationSideThreshold These above two configs decide if insert bloom filter for the join, which take effect in Spark java code.

So they are transparent to Gluten, right?

[github-actions]

Run Gluten Clickhouse CI on x86

[github-actions]

Run Gluten Clickhouse CI on x86

[github-actions]

Run Gluten Clickhouse CI on x86

[github-actions]

Run Gluten Clickhouse CI on x86

[github-actions]

Run Gluten Clickhouse CI on x86