【Pandas】检查是否有空值、处理空值

深藏阁楼爱情的钟 2023-07-12 11:39 4阅读 0赞

#### 1.创建有空值的DataFrame ####

import numpy as np
    import pandas as pd
    
    dates = pd.date_range("20200307", periods=4)
    df1 = pd.DataFrame(np.arange(12).reshape(4, 3), index=dates, columns=["A", "B", "C"])
    df2 = pd.DataFrame(df1, index=dates, columns=["A", "B", "C", "D"])  # 新增D列，却不赋值，NaN表示空值
    print(df2)
    # 打印输出：
    # A B C D
    # 2020-03-07 0 1 2 NaN
    # 2020-03-08 3 4 5 NaN
    # 2020-03-09 6 7 8 NaN
    # 2020-03-10 9 10 11 NaN

#### 2.检查是否有空值 ####

print(df2.isnull())  # 是空值返回True，否则返回False
    print(np.any(df2.isnull()))  # 只要有一个空值便会返回True，否则返回False
    print(np.all(df2.isnull()))  # 全部值都是空值便会返回True，否则返回False
    # 输出结果：
    # A B C D
    # 2020-03-07 False False False True
    # 2020-03-08 False False False True
    # 2020-03-09 False False False True
    # 2020-03-10 False False False True
    # True
    # False

#### 3.给NaN赋值 ####

df2.iloc[0, 3] = 10  # 直接给某个位置赋值
    print(df2)
    # 打印输出：
    # A B C D
    # 2020-03-07 0 1 2 10.0
    # 2020-03-08 3 4 5 NaN
    # 2020-03-09 6 7 8 NaN
    # 2020-03-10 9 10 11 NaN
    
    series = pd.Series([11, 12, 13], index=dates[1:4])
    df2["D"] = series  # 同时给D列赋多个值
    print(df2)
    # 打印输出：
    # A B C D
    # 2020-03-07 0 1 2 NaN
    # 2020-03-08 3 4 5 11.0
    # 2020-03-09 6 7 8 12.0
    # 2020-03-10 9 10 11 13.0

#### 4.去除有空值的行或列 ####

df2.loc["2020-03-10", ["A", "B", "C"]] = [11, 12, 15]
    df2.fillna("null")  # 把空值填充成null
    
    # dropna(axis,how,subset)方法会删除有空值的行或列，
    # axis为0是行，axis为1是列,
    # how为any时该行或列只要有一个空值就会删除，all是全都是空值才删除
    # subset是一个列表，指定某些列
    df2.dropna(axis=0, how="any", subset=["A", "D"])