rle_array.autoconversion module

auto_convert_to_rle(df: pandas.core.frame.DataFrame, threshold: Optional[float] = None) → pandas.core.frame.DataFrame

Auto-convert given DataFrame to RLE compressed DataFrame.

Important

Datetime columns are currently not compressed due to pandas not supporting them.

Please note that RLE can, under some circumstances, require MORE memory than the uncompressed data. It is not advisable to set threshold to a value larger than 1 except for testing purposes.

Parameters

  • df – Input DataFrame, may already contain RLE columns. This input data MIGHT not be copied!

  • threshold –

    Compression threshold, e.g.:

    • None: compress all

    • 1.0 compresses only if RLE does NOT take up more space

    • 0.5 compresses if at least 50% memory are safed

    • 0.0 do not compress at all

Raises

ValueError – If threshold is negative.

decompress(df: pandas.core.frame.DataFrame) → pandas.core.frame.DataFrame

Decompress all RLE columns in the provided DataFrame.

Parameters

df – Input DataFrame. This input data MIGHT not be copied!