def f():
    df = pd.get_dummies(column_values, columns=input_columns_names)
    df2 = None
    for column in input_columns_names:
        for input_column_value in query_table[column]:
            if str(input_column_value).lower()=="nan":
                continue
            if(isinstance(input_column_value, int)):
                input_column_value=str(int(input_column_value))
            else:
                input_column_value=str(input_column_value)
            col1 = pd.DataFrame(df[str(column) + "_" + input_column_value])
            col1.rename(columns={col1.columns[0]: str(column) + "_" + str(input_column_value)})
            if df2 is None:
                df2 = col1
            else:
                df2 = pd.concat([df2, col1], axis=1)
    df = pd.DataFrame.empty
    correlation = df2.corr()
    for output_column in correlation:
        for unchecked_column in unchecked_column_names:
            if output_column.find(unchecked_column+"_") != -1:
                correlation.drop(output_column, inplace=True, axis=1)
    for output_column in correlation.index:
        for checked_column in checked_column_names:
            if output_column.find(checked_column + "_") != -1:
                correlation.drop(output_column, inplace=True, axis='index')
    return correlation.abs()