I have to final tables, table_1 and table_2, first one with 6.4B rows and second one with 6.3B rows. None of this tables are clustered

On top of this tables, each one have a stream, standard ones, stream_1 and stream_2 respectively.

When i query those streams, grouping by a date column, first one response in seconds,

query profile like this,

Indicating that catch 8 million new rows approx.

But when i query second one, takes over an hour and the query profile indicates this:

Indicating that makes like to table scans and then joining them, i don’t get it.

Any thoughts on why is doing this?