Sparse pd.DataFrame

df_sparse.py

data = [
    ['User', 'Content Type', 'File', 'Mb'],
    ['User 1', 'Text', 'F1', 0.1],
    ['User 1', 'Text', 'F2', 0.5],
    ['User 1', 'Video', 'V1', 100],
    ['User 2', 'Text', 'F1', 2],
    ['User 2', 'Video', 'V1', 75],
    ['User 3', 'Video', 'V2', 60],
]

df = pd.DataFrame(data[1:], columns=data[0])


class DFSparser:
    def __init__(self, df, group_by_column):
        self.df = df
        self.group_by_column = group_by_column
    
    def get_mask(self):
        mask = pd.concat(
            (
                pd.concat([
                    group[1].duplicated()
                    for group in self.df.groupby(self.group_by_column)[c]
                ])
                for c in self.df.columns.tolist()
            ),
            axis=1,
        )
        return mask


    def apply(self, fill=None):
        out = self.df.where(~self.get_mask())
        if fill is not None:
            return out.fillna(fill)
        return out