Python 数据清洗删除缺失值替换缺失值详情

df.isnull()

df.isnull().sum()

df.isnull().sum(axis='columns')

df.isnull().all()

df1 = df.drop(columns='植物园')
df1 = df.drop('植物园', axis=1)

df1 = df.loc[:, ~(df.isnull().all())]

df1[~(df1.isnull().any('columns'))]

df.dropna(axis='columns', how='all', inplace=True)

df.dropna(axis='index', how='any', inplace=True)

df.fillna(method='bfill')

df.fillna(method='bfill')

df.duplicated()

df.drop_duplicates()

df.duplicated(subset=['天坛'], keep=False)

df.drop_duplicates(subset=['天坛'], keep=False)

df.replace(np.nan, -9999)

df.replace([np.nan, 0], -9999)

def aqi_level(aqi):
    if aqi>0 and aqi<=50:
        level = '优'
    elif aqi>50 and aqi<=100:
        level = '良'
    elif aqi>100 and aqi<=150:
        level = '轻度污染'
    elif aqi>150 and aqi<=200:
        level = '中度污染'
    elif aqi>200 and aqi<=300:
        level = '重度污染'
    else:
        level = '严重污染'
    return level

# 数据预处理一下  将所有的类型都转为 AQI
aqi = df[df['type']=='AQI']

aqi['东四'].apply(aqi_level)

aqi.loc[:, ~aqi.columns.isin(['date', 'hour', 'type'])].apply(np.mean, axis=0)
aqi.loc[:, ~aqi.columns.isin(['date', 'hour', 'type'])].applymap(aqi_level)