pandas replacen内部使用函数

在 Pandas 中，`replace()` 方法内部使用了 `_get_mapper()` 函数来获取一个字典映射，将要替换的值映射为新的值。`_get_mapper()` 函数的源代码如下：

def _get_mapper(self, mapper=None):
    if callable(mapper):
        mapper = mapper(self)

    if isinstance(mapper, (dict, Series)):
        if isinstance(mapper, dict):
            keys = list(mapper.keys())
            if any(k is None or com.is_datetime64tz(k) for k in keys):
                keys = [com.pandas_ify(k) for k in keys]
            values = [mapper[k] for k in keys]
        else:
            keys = mapper.index
            values = mapper.values
        return dict(zip(keys, values))

    # if a list of tuples, convert to dict
    elif isinstance(mapper, (list, tuple)):
        if not all(isinstance(x, (tuple, list)) and len(x) == 2 for x in mapper):
            raise ValueError("If 'mapper' is a list or tuple, it should be a list/tuple of tuples")
        return dict((k, v) for k, v in mapper)

    elif mapper is None:
        return {}

    else:
        raise TypeError("If 'mapper' is None, it is expected to be a dict, list or tuple, not {0}".format(type(mapper)))

该函数的作用是将传入的 `mapper` 对象（可以是字典、Series、列表或元组）转换为一个字典映射，如果 `mapper` 是一个可调用对象，则调用该对象并传入当前 DataFrame/Series 对象作为参数。

`replace()` 方法将传入的 `to_replace` 参数和 `value` 参数都作为 `mapper` 参数传入 `_get_mapper()` 函数，然后将返回的字典映射应用到 DataFrame/Series 中，实现值的替换。