Longer term patch to replace the workaround in #45 (which should be merged in the mean time)
My preference here is to enforce a schema for each dataset when constructing the polars df (i.e. provide a schema kwarg to pl.from_dicts). I think we should get the schema from the dataset's own column metadata on data.cdc.gov (e.g. for NHSN hrd prelim: data.cdc.gov/api/views/mpgq-jmmr.json; see stackoverflow.com/questions/24745820/socrata-get-data-types-of-dataset-columns).
Longer term patch to replace the workaround in #45 (which should be merged in the mean time)
My preference here is to enforce a schema for each dataset when constructing the polars df (i.e. provide a
schemakwarg topl.from_dicts). I think we should get the schema from the dataset's own column metadata on data.cdc.gov (e.g. for NHSN hrd prelim: data.cdc.gov/api/views/mpgq-jmmr.json; see stackoverflow.com/questions/24745820/socrata-get-data-types-of-dataset-columns).