Add APPROX_TOP_COUNT function

[david542542]

BigQuery has a nice feature/function called APPROX_TOP_COUNT that returns a struct of the top N values with their approximate count. This is similar to the DuckDB SUMMARIZE command, but allows a column (or other arbitrary expression). I think this would be a great function/feature to add if this is possible. Postgres also has most_common_vals within pg_stats.

[carlopi]

Hi! Thanks, looks a cool feature request, can you maybe move this to main duckdb's repo discussions?

Thanks!

[soerenwolfers]

since implemented as approx_top_k duckdb/duckdb#12653

[carlopi]

Thanks, indeed this will be available in a month or so with duckdb release 1.2