#coding=utf-8
import pandas as pd
import numpy as np

data={
    'company':['B','A','B',np.nan,'C'],
    'gender':['female','female','male','male','female'],
    'num':[40,31,28,28,50]
}

df=pd.DataFrame(data)
print(df)
'''
  company  gender  num
0       B  female   40
1       A  female   31
2       B    male   28
3     NaN    male   28
4       C  female   50
'''




#举例1、value_counts() 统计分类变量中每个类的数量，注：指定的列不能是数值的列
print(df.gender.value_counts())
'''
female    3
male      2
Name: gender, dtype: int64
'''
print(df.company.value_counts())
'''
B    2
C    1
A    1
Name: company, dtype: int64
'''
print(df.gender.value_counts(normalize=True))
'''
解释：normalize=True表示返回占比的百分比

female    0.6
male      0.4
Name: gender, dtype: float64
'''





#举例2、isna(),判断是否有缺失值，有缺失值返回True，无缺失值返回false
print(df.isna())
'''
   company  gender    num
0    False   False  False
1    False   False  False
2    False   False  False
3     True   False  False
4    False   False  False
'''
print(df.company.isna())
'''
0    False
1    False
2    False
3     True
4    False
Name: company, dtype: bool
'''
print(df.gender.isna())
'''
0    False
1    False
2    False
3    False
4    False
Name: gender, dtype: bool
'''
print(df.loc[df.company.isna()])
'''
  company gender  num
3     NaN   male   28
'''





#举例3、any()
#大多数情况下数据量比较大，使用isna()效率太低下，any()和isna()综合使用可以判断某一列是否含有缺失值df.isna().any()
print(df.isna().any())
'''
company     True
gender     False
num        False
dtype: bool
'''





#举例4、dropna()，删除缺失值
print(df.dropna())
'''
注：未传入参数inplace=True，并未真正删除原df对象的缺失值
  company  gender  num
0       B  female   40
1       A  female   31
2       B    male   28
4       C  female   50
'''





#举例5、fillna(value,method)，填充缺失值
'''
参数解释：
value=? 指定缺失值等于某值
method= pad/ffill   backfill/bfill
       pad/ffill      表示用上一个值填充该缺失值
       backfill/bfill 表示用下一个值填充该缺失值
'''
print(df.fillna(value=0))
'''
  company  gender  num
0       B  female   40
1       A  female   31
2       B    male   28
3       0    male   28
4       C  female   50

可以看到缺失值被0所替换
注：未传入inplace=True，原df对象并未被真正修改
'''

print(df.fillna(method="pad"))
'''  
或print(df.fillna(method="ffill"))
    company  gender  num
0       B  female   40
1       A  female   31
2       B    male   28
3       B    male   28
4       C  female   50

可以看到缺失值被上一个值"B"所替换
'''

print(df.fillna(method="backfill"))
'''
  company  gender  num
0       B  female   40
1       A  female   31
2       B    male   28
3       C    male   28
4       C  female   50

可以看到缺失值被下一个值"C"所替换
'''





#举例6、sort_values(by,ascending)，排序
print(df.sort_values(by="num"))
'''
  company  gender  num
2       B    male   28
3     NaN    male   28
1       A  female   31
0       B  female   40
4       C  female   50
'''
print(df.sort_values(by="num",ascending=False))
'''
ascending=False 控制降序排序，默认为True升序排序

  company  gender  num
4       C  female   50
0       B  female   40
1       A  female   31
2       B    male   28
3     NaN    male   28
'''