Proposed different CSV column selection logic

[JensRantil]

I have an interest in being able to parse additional transaction metadata from the CSV file, namely:

• An auxiliary date
• A code.
• A transaction state.
• A balance assertion.

Currently the CSV parsing logic has few issues:

• The actual columns used in the CSV files is not really well-documented; Neither README, nor website, states what goes in which column.
• There is no logic to parse CSV columns in a different CSV file format than the assumed input. My bank allows me to download XLS files from them - it would be great if I simply could convert that Excel file into CSV and not have to transform the CSV file into a format that into-ledger can handle.
• There is no logic to parse optional columns.

Proposal: I propose that

• the -ic flag is removed.
• every CSV-parseable transaction field gets an equivalent flag for which CSV column that field is stored in.
  • Required transaction fields (date, amount, description) default to >=0 index equivalent to today's functionality.
  • Optional fields (which I intend to add) default to field index <0. Enable by setting to a >=0 index.

What do you think about this? I'd be willing to submit a PR for this.

[JensRantil]

Let me know if you'd like me to split this issue into two; One that deals with the new CSV field logic, and the second one about adding support for the additional transaction fields I want.