Partition indicesΒΆ


A secondary index is persisted as a Parquet file with the following (Parquet) schema: The fieldname corresponds to the name of the column in the persisted DataFrame. The partition is a list of partition identifiers, as used in the keys of the partitions map and the data filename. (Note: the partition identifier is used instead of the data filename as a single partition can span multiple files containing different column sets using the same row selection.)