pandas_utils-0.1.1.tar.gz

import pandas as pd from typing import Tuple, List __all__ = ['groupcount_differences_between_rows'] def groupcount_differences_between_rows( df1: pd.DataFrame, df2: pd.DataFrame ) -> pd.DataFrame: """ Group differences between rows and count entries of each group. Example: Input: df1: df2: A B C A B C ---------- ---------- 1 1 1 1 1 1 1 1 1 1 99 1 1 1 1 1 99 1 1 1 1 1 1 99 1 1 1 99 99 1 Output: group count indices --------------------------- [B] 2 [1, 2] [C] 1 [3] [B, C] 1 [4] For more examples see the tests of this function. """ df1, df2 = check_and_parse_dataframes(df1, df2) diffs = pd.DataFrame(columns=["group", "count", "indices"]) for (i, row1), (_, row2) in zip(df1.iterrows(), df2.iterrows()): group = get_names_of_unequal_fields(row1, row2) if group: diffs = diffs.append(dict( group=tuple(group), count=1, indices=[i] ), ignore_index=True) diffs = (diffs.groupby("group", as_index=False) .agg({'count': 'sum', 'indices': 'sum'})) return diffs def get_names_of_unequal_fields( row1: pd.Series, row2: pd.Series ) -> List[str]: return [str(name) for (name, field1), (_, field2) in zip(row1.items(), row2.items()) if field1 != field2] def check_and_parse_dataframes( df1: pd.DataFrame, df2: pd.DataFrame ) -> Tuple[pd.DataFrame]: if df1.shape != df2.shape: raise ValueError(f'df1 and df2 do not have same shape: {df1.shape}, {df2.shape}') for df in (df1, df2): if len(df.columns) != len(set(df.columns)): raise ValueError(f"Dataframes are not allowed to have non-unique columns: {df1.columns}, {df2.columns}") if set(df1.columns) != set(df2.columns): raise ValueError(f'df1 and df2 do not have the same set of columns: {df1.columns}, {df2.columns}') if any(list(df1.columns != df2.columns)): df2 = df2[df1.columns] if any(list(df1.dtypes != df2.dtypes)): raise TypeError(f'Datatypes of same named columns in df1 and df2 do not match:') return df1, df2 if __name__ == '__main__': df1 = pd.DataFrame(dict( A=[1, 1, 1], B=[1, 1, 1], C=[1, 1, 1] )) df2 = pd.DataFrame(dict( A=[1, 9, 9], B=[9, 1, 1], C=[9, 9, 9] )) observation = groupcount_differences_between_rows(df1, df2) diffs2 = pd.DataFrame(dict( group=[("Comments",), ("Comments",), ("Old_Price_EUR",), ("Old_Price_EUR",)], count=[1, 1, 1, 1], indices=[[0], [1], [5], [6]] ))