ParquetDataCatalog write_data() no longer supports base_templates for filenaming conventions

[sdk451]

Bug Report

code that previously worked under 1.218 e.g.

catalog = ParquetDataCatalog(CATALOG_PATH)
catalog.write_data(bars, basename_template=f'{filedate}'+'-part-{i}')

now fails as basename_template has been removed as an optional var to write_data().

Being able to version and control naming conventions for files written to to data catalog is an important feature that should be retained.

[cjdsellers]

Hi @sdk451

Thanks for reaching out.

The basename_template was removed in favor of adopting a fixed naming convention for files that includes a human-readable timestamp for the start and end of the time range. This allowed us to introduce additional features such as on-the-fly downloads and consolidations.

Unfortunately, with the fast pace of development there was not a formal deprecation strategy.

[sdk451]

Ok. I did have a look at the documentation and I couldn't see anything about the breaking change or how to manage catalog filenaming under 1.219. What's the approved approach for setting filenames into the catalog now? (Apologies I can't post on discord anymore - my discord account won't allow me to add my mobile).

[faysou]

You can do catalog.reset_file_name() then not worry about file names anymore.