Request: Allow deserializing missing fields

I have some TSV files that have trimmed rows to avoid unnecessary `\t\t\t` chains.

It would be nice if this library could support an option to return default value for these missing fields.

In my case I don't even need those missing fields, I just need first two, but I get `expected field, but got end of row` for some fields that I am not even interested in.

Thx for the crate, anyway ;)

Start by locating the deserialization path that emits “expected field, but got end of row” and read the surrounding tests. Determine how an option for trimmed TSV rows should supply defaults for missing fields while allowing callers to read only the fields they need. Done means the behavior is covered by tests for missing trailing fields and normal rows.

Written by the indexing model from the issue text.