Add script to identify candidate changes to column maps

[krivard]

Overview

Quarterly updates for sources that frequently change the layout/schema/spelling of their raw files are tedious. The 2025Q1 update for EIA 930 dropped 9 and added 30 columns.

A script that compares the new source file(s) to the existing maps and summarizes changes would help immensely.

Success Criteria

A good script would output:

• columns in the map that are not in the new source file (to identify deleted columns)
• columns in the new source file that are not mapped (to identify added columns)
• column headers in the new source file paired with their PUDL column names (to identify changes to column order)

[e-belfer]

Right now the extraction step spits out any unmapped and falsely mapped columns as part of an error. I'm preferential to this as a debugging output rather than a script we need to remember to maintain - is there a component of what you're proposing that isn't addressed by this current workflow, or a reason you'd prefer to pull it into a script?

[krivard]

The last time I had to update column maps was for the Q2 release, and at that time it absolutely did not give me helpful messages about it. Was the error added more recently?

[e-belfer]

The validate method in the GenericExtractor() should be where this is happening, it's possible there's a bug here we haven't caught if this wasn't catching errors when you were working on the data?