querycolumn

Simple extension to Pandas that makes it easier to select columns in (very) wide DataFrames. If you name your columns in a hierarchical fashion with a separator (e.g, as you might get from pd.normalize_json(), it lets you select a column or group of columns easily and with tab completion.

Here's a quick demo of what this looks like.

    >>> import pandas as pd
    >>> from querycolumn import patch_dataframe_with_query_columns
    >>> patch_dataframe_with_query_columns()
    >>> df = pd.DataFrame([{'person.size.height': 3, 'person.size.weight': 4}])
    >>> df.qc.person
    <QueryColumns @ person: size>
    >>> df.qc.person.size
    <QueryColumns @ person.size: height, weight>
    >>> df.qc.person.size.weight
    0    4
    Name: person.size.weight, dtype: int64
    >>> df[df.qc.person.size]
    >>> 

QueryColumns does its magic by patching a Descriptor into the DataFrame class. When you retrieve the attribute qc from a frame, `QueryColumns.get()' is invoked; this is when QueryColumns introspects its parent dataframe and returns a magical object with tab completion for segmented column names.

Note I wrote QueryColumns after working with Pandas for about a week. I don't know if it encourages elegant use of the library, but I find it useful for my usecase. I'm practically certain that it doesn't extend Pandas as one should (there are explicit APIs for that), but I couldn't think of a way to do it without the Descritor protocol.