value_counter.py 文件源码

def fill_histogram(self, idf, columns):
        """Fill input histogram with column(s) of input dataframe

        :param idf: input data frame used for filling histogram
        :param list columns: histogram column(s)
        """

        name = ':'.join(columns)
        if name not in self._counts:
            # create an (empty) value counts dict
            self._counts[name] = Counter()
        # value_counts() is faster than groupby().size(), but only works for series (1d).
        # else use groupby() for multi-dimensions
        g = idf.groupby(by=columns).size() if len(columns) > 1 else idf[columns[0]].value_counts()
        counts = Counter(g.to_dict())
        # remove specific keys from histogram before merging, if so requested
        counts = self.drop_requested_keys(name, counts)
        self._counts[name].update(counts)