The default required column should also be a default list of the column_filters

[chinyeungli]

I keep getting the following error

CRITICAL: Row 0 is missing required values for columns: about_file_path, name

with the following transform config file:

required_columns:
    - license_expression
    
column_filters:
    - license_expression

Then I realize if I add the name and the about_file_path in the column_filters list, everything went fine.

Since these two fields are required by default, user should not need to put these 2 fields in the column_filters list.

[pombredanne]

The latest code in the ranch has been updated and the column_filters has been renamed to kept_columns.

Note that per #345 name should not be mandatory , so this live only the about_file_path field.

We would need to discuss this about_file_path field as I explained in #364 ... IMHO this should not be needed with gen as the placement of a generated .ABOUT file should be relative to the documented about_resource and therefore having an explicit about_file_path is a likely source of problems and confusion. In most cases the path would be the about_resource path with a .ABOUT extension.

[chinyeungli]

@pombredanne I don't see any kept_columns in the latest develop branch nor 361-make-transform-work... has the latest code been pushed/merged?

[pombredanne]

@chinyeungli you can find the new kept columns in the 361-make-transform-work branch... See https://github.com/nexB/aboutcode-toolkit/blob/361-make-transform-work/src/attributecode/transform.py#L123
May be you are not up to date?

[chinyeungli]

Since there is API changes, I mark this for v5 ticket.

[chinyeungli]

Not going to fix as we already have a detailed doc about what need to be included:

* field_filters:
An optional list of field names that should be kept in the transformed CSV/JSON. If
this list is provided, all the fields from the source CSV/JSON that should be kept
in the target CSV/JSON must be listed regardless of  either standard or required
fields. If this list is not provided, all source CSV/JSON fields are kept in the
transformed target CSV/JSON.