Is hydra columnar dropped?

[ivan-v-kush]

What's wrong?

It seems development of hydra columnar is discontinued. New PRs is not reviewed, not merged.

[SebastienGllmt]

For anybody who finds this issue, it seems they've moved on to instead develop pg_duckdb as commits stop to this repo at the same time as the Hydra team helped ship the 0.3.0 version of pg_duckdb

[digEmAll]

Correct me if I'm wrong, but pg_duckdb is not the same as hydra.

AFAIK pg_duckdb embeds the duckdb analytic query capabilities inside a normal (i.e. non-columnar) postgres.
So basically you still store data inside postgres row-wise tables, but you query them with duckdb engine, gaining some speed on analytic queries; furthermore you can directly query parquet file, csv etc.

hydra-db instead adds to postgres the columnar storage for tables, which make postgres much faster for analytics.

But, since in pg_duckdb the data is still stored row-wise (unless you use parquet file of course), I think hydradb should perform better... am I right ?