i've got the following situation:
i have downloaded a dataset for a yellow taxis trips (https://s3.amazonaws.com/nyc-tlc/trip+data/yellow_tripdata_2020-01.csv)
i have imported it as a new entity, with structure defined by sampling and loaded it.
i have discovered, that sample that i have defined incorrectly defined one of the amounts (tolls_amount field is booleans instead of double)
at this point i redefined it in the source section (external metadata), removed all loads and rerun the load, the internal metadata in discovery section is fine, i can see the disctribution but, i cannot use the query.
it seems that the definition of the postgres foreign table, that maps the file - is still using the boolean as the data type.
what is the safest way to redefine that definition? at which point it can be regenerated?