QA had detected high memory usage on some columns when executing the report downloader (e.g. https://github.com/project-koku/koku/blob/46cd8b2f0626a9fa0012f7a835b5a5a6d980be4b/koku/masu/external/downloader/aws/aws_report_downloader.py#L121).

The idea to solve that is to change the str data type to StringDtype, which probably would consume less memory, as well as follows Pandas best practices (https://pandas.pydata.org/pandas-docs/stable/user_guide/basics.html#dtypes).

It is needed to perform this action in all _report_downloader.py files, including the _local_report_downloder.py, to make it consistent.