Fix sql bugs (#7219)

* fix nil when explaining

* add plain details when running full scan

* skip files by timestamp

* skip file by timestamp

* refactor

* handle filter by time

* skip broker memory only if it has unflushed messages

* refactoring

* refactor

* address comments

* address comments

* filter by parquet stats

* simplify

* refactor

* prune old code

* optimize

* Update aggregations.go

* ensure non-time predicates are properly detected

* add stmt to populatePlanFileDetails

This helper function is a great way to centralize logic for populating file details. However, it's missing an optimization that is present in executeSelectStatementWithBrokerStats: pruning Parquet files based on column statistics from the WHERE clause.

Aggregation queries that fall back to the slow path could benefit from this optimization. Consider modifying the function signature to accept the *SelectStatement and adding the column statistics pruning logic here, similar to how it's done in executeSelectStatementWithBrokerStats.

* refactoring to work with *schema_pb.Value directly after the initial conversion

0 files changed, 0 insertions, 0 deletions