We have a mapping, so if the headers of the cost centre csv are changed, we only have to update them in the csv_updater script, and not in other places in the codebase.
The file uploaded to S3 now only contains the fields used by this app, instead of everything in the source csv.
We have a mapping, so if the headers of the cost centre csv are changed, we only have to update them in the csv_updater script, and not in other places in the codebase.
The file uploaded to S3 now only contains the fields used by this app, instead of everything in the source csv.