Meltano

Hey,
using tap-s3-csv.
Problem: foreign key column contains NULL values. It is loaded as VARCHAR and then it cannot be used as foreign key, because NULL values are loaded as empty VARCHAR `''`, not NULL value.
There is override for dates property in config, but not for INT.
Any solution exists?

<https://github.com/s7clarke10/pipelinewise-tap-s3-csv|Steve Clarke's ><https://github.com/s7clarke10/pipelinewise-tap-s3-csv|fork> has a `set_empty_values_null` setting: When set true will emit `null` (the JSON equivalent of None) instead of an empty string.

Thanks. But still, I would expect better autodetection of column data types. Take DuckDB's read_csv_auto(), for instance :wink:

One more issue - I have CSV files without a column(s) which represent primary key.
I want to do full-refresh every time, they are small files.
But `key_properties` seem to be mandatory.
Ha, not for all targets it seems! The extract/load passed against Postgres, but failed against Snowflake! Strange.
Anyway, I can specify all columns in `key_properties` as a workaround

There might be a setting to disable the PK requirement