Question

我有一个包含单个列的大数据，其中每一行的格式为：

82283343~Electronics~Mobile Cases & Covers

我想将上面的列在波浪线分成三列（82283343，Electronics，Mobile Case & Covers）。我写了以下代码：

df= df._id.map(lambda x: x.split('~'))

但这根本没有效率，我最终关闭了终端。还有更好的方法吗？

Answer 1

我尝试做一些测试并选择最佳方法。

最快的一个是从列_id创建列表并由原生python split('~')拆分：

df[['one', 'two', 'three']] = pd.DataFrame([ x.split('~') for x in df['_id'].tolist() ])

import pandas as pd

#test list
x =['82283344~Electronics~Mobile Cases & Covers', '82283346~Electronics~Mobile Cases & Covers', '82283343~Electronics~Mobile Cases & Covers']
#100000 lists
x = x * 100000

#create new df with column _id
df = pd.DataFrame({'_id': x })

print df.head()
                                         _id
0  82283344~Electronics~Mobile Cases & Covers
1  82283346~Electronics~Mobile Cases & Covers
2  82283343~Electronics~Mobile Cases & Covers
3  82283344~Electronics~Mobile Cases & Covers
4  82283346~Electronics~Mobile Cases & Covers

def DF(df):
    df[['one', 'two', 'three']] = pd.DataFrame([ x.split('~') for x in df['_id'].tolist() ])

def AP(df):
    df['one'] = df._id.apply(lambda x: x.split('~')[0])  
    df['two'] = df._id.apply(lambda x: x.split('~')[1])
    df['three'] = df._id.apply(lambda x: x.split('~')[2])

def EX(df):
    df[['one', 'two', 'three']] = df._id.str.split('~', expand=True)

def SP(df):
    df['one'] = df['_id'].str.split('~').str[0]
    df['two'] = df['_id'].str.split('~').str[1]
    df['three'] = df['_id'].str.split('~').str[2] 

DF(df)
print df.head()
AP(df)
print df.head()
EX(df)
print df.head()
SP(df)
print df.head()

重复4次：

                                          _id       one          two  \
0  82283344~Electronics~Mobile Cases & Covers  82283344  Electronics   
1  82283346~Electronics~Mobile Cases & Covers  82283346  Electronics   
2  82283343~Electronics~Mobile Cases & Covers  82283343  Electronics   
3  82283344~Electronics~Mobile Cases & Covers  82283344  Electronics   
4  82283346~Electronics~Mobile Cases & Covers  82283346  Electronics   

                   three  
0  Mobile Cases & Covers  
1  Mobile Cases & Covers  
2  Mobile Cases & Covers  
3  Mobile Cases & Covers  
4  Mobile Cases & Covers

定时：

In [125]: %timeit DF(df)
     ...: %timeit AP(df)
     ...: %timeit EX(df)
     ...: %timeit SP(df)
     ...: 
1 loops, best of 3: 332 ms per loop
1 loops, best of 3: 564 ms per loop
1 loops, best of 3: 668 ms per loop
1 loops, best of 3: 1.09 s per loop

Answer 2

我认为将一列分为三列并再次保存到同一数据帧应该可以正常工作：

df = df['_id'].str.split('~', 3, expand=True)

请尝试此操作，如果您有任何问题，请告诉我们。

如何在pandas（python）中将单个列拆分为三列？

2 个答案: