Save statistics for all columns (Including Non-Indexed Columns) in TimescaleDB

[pantonis]

What type of enhancement is this?

Performance

What subsystems and features will be improved?

Query executor, Query planner

What does the enhancement do?

It would be highly beneficial if TimescaleDB could save statistics for all columns, not just indexed ones, for compressed tables. This approach would function similarly to how Parquet files store metadata and column-level statistics.
Saving statistics for all columns would optimize the elimination of chunks, making query execution more efficient. It will also provide more comprehensive statistics would allow predicates to work more effectively, especially for filtering and range queries on non-indexed columns

Parquet files leverage column-level statistics to great effect, leading to significant performance gains. Implementing a similar mechanism in TimescaleDB would provide comparable benefits.
This feature would be a great enhancement for workloads that rely on frequent queries with non-indexed predicates, particularly in analytics where we have ad-hoc queries all the time

Implementation challenges

No response