打卡41天

pandas分组：

groupby方法

按指定列分组，返回每个组对应数的索引

data = {
    'name':['zhangsan','lisi','wangwu','liliu','tom','john'],
    'class':['one','two','three','one','two','three'],
    'score':[54,56,57,58,59,80]
}
df = pd.DataFrame(data)
print(df)

'''
       name  class  score
0  zhangsan    one     54
1      lisi    two     56
2    wangwu  three     57
3     liliu    one     58
4       tom    two     59
5      john  three     80
'''

print(df.groupby('class').groups)

'''
{'one': [0, 3], 'three': [2, 5], 'two': [1, 4]}
'''

df.agg()方法对分组聚合

print(df.groupby('class').agg(np.mean))

'''
       score
class       
one     56.0
three   68.5
two     57.5
'''

transform方法，与原数据形状一直的聚合数据

df['score_percent'] = df.groupby('class')['score'].transform(np.sum)
df['score_percent'] = df['score']/df['score_percent']
print(df)

'''
       name  class  score  score_percent
0  zhangsan    one     54       0.482143
1      lisi    two     56       0.486957
2    wangwu  three     57       0.416058
3     liliu    one     58       0.517857
4       tom    two     59       0.513043
5      john  three     80       0.583942
'''

pandas连接合并

pandas数据帧的连接，与数据库中表的连接方式类似

水平连接

merge方法：on参数来设置连接条件

data = {
    'name':['zhangsan','lisi','wangwu','liliu','tom','john'],
    'class':['one','two','three','one','two','three'],
    'score':[54,56,57,58,59,80]
}
df = pd.DataFrame(data)

data1 = {
    'name':['zhangsan','lisi','wangwu','liliu','tom','john'],
    'math_score':[80,80,87,56,90,70]
}
df1 = pd.DataFrame(data1)
#将df与df1连接，连接条件'name'
df = df.merge(df1,on='name')
print(df)

'''
       name  class  score  score_percent  math_score
0  zhangsan    one     54       0.482143          80
1      lisi    two     56       0.486957          80
2    wangwu  three     57       0.416058          87
3     liliu    one     58       0.517857          56
4       tom    two     59       0.513043          90
5      john  three     80       0.583942          70
'''

通过merge方法中的how参数，'left','right','outer'，设置左连接，右连接，外连接

垂直连接：

concat方法

df = pd.concat([df,df])
print(df)

'''
       name  class  score  score_percent  math_score
0  zhangsan    one     54       0.482143          80
1      lisi    two     56       0.486957          80
2    wangwu  three     57       0.416058          87
3     liliu    one     58       0.517857          56
4       tom    two     59       0.513043          90
5      john  three     80       0.583942          70
0  zhangsan    one     54       0.482143          80
1      lisi    two     56       0.486957          80
2    wangwu  three     57       0.416058          87
3     liliu    one     58       0.517857          56
4       tom    two     59       0.513043          90
5      john  three     80       0.583942          70
'''