pandas.core.window.rolling.Rolling.count #

滚动。计数( numeric_only = False ) [来源] #

计算非 NaN 观测值的滚动计数。

参数
numeric_only布尔值,默认 False

仅包含 float、int、boolean 列。

1.5.0 版本中的新增内容。

返回
系列或数据框

返回类型与具有 dtype 的原始对象相同np.float64

也可以看看

pandas.Series.rolling

使用系列数据进行滚动调用。

pandas.DataFrame.rolling

使用 DataFrame 调用滚动。

pandas.Series.count

系列的聚合计数。

pandas.DataFrame.count

聚合 DataFrame 的计数。

例子

>>> s = pd.Series([2, 3, np.nan, 10])
>>> s.rolling(2).count()
0    NaN
1    2.0
2    1.0
3    1.0
dtype: float64
>>> s.rolling(3).count()
0    NaN
1    NaN
2    2.0
3    2.0
dtype: float64
>>> s.rolling(4).count()
0    NaN
1    NaN
2    NaN
3    3.0
dtype: float64