Add top level functions to_protobuf and from_protobuf which simplify column conversion without the user having to create a MessageConverter.
Add top level functions df_to_protobuf and df_from_protobuf which can be passed a dataframe without a MessageConverter. These functions also have expanded: bool args which handle the expansion and contraction of data from/to the struct which is the output/input of the encoding/decoding udfs. The functions call into new MessageConverter.df_to/from_protobuf methods.
Also, since spark already uses flatten nomenclature, which is different from what we are doing here, we change our documentation to use the work expand or contract to describe that we are unpacking or packing a struct, respectively.
Add top level functions
to_protobufandfrom_protobufwhich simplify column conversion without the user having to create aMessageConverter.Add top level functions
df_to_protobufanddf_from_protobufwhich can be passed a dataframe without aMessageConverter. These functions also haveexpanded: boolargs which handle the expansion and contraction of data from/to the struct which is the output/input of the encoding/decodingudfs. The functions call into newMessageConverter.df_to/from_protobufmethods.Also, since spark already uses
flattennomenclature, which is different from what we are doing here, we change our documentation to use the work expand or contract to describe that we are unpacking or packing a struct, respectively.