Calling batch_slice_columns multiple times results in unexpected behavior

[clinejj]

Given a file with the headers:
supplier,sku,name,description,shipping,price,quantity_unit,minimum_quantity,maximum_quantity

After running in a before_batch_import callback,
importer.batch_slice_columns(%w[supplier sku price minimum_quantity maximum_quantity])
the headers are sliced as expected.

However, if I wanted to further reduce the results set, calling
importer.batch_slice_columns(%w[sku price minimum_quantity maximum_quantity])
does not reduce the csv_lines as expected due to @use_indexes here referring to the original indices but not the resulting indices.

This ends up replacing the values unexpectedly (since it uses the indices in @use_indexes which may no longer apply to the sliced results).

Ideally, running batch_slice_columns multiple times would continually reduce the column set based on the results of the prior call.