从字符串列 python 中提取 substring

Question

I have a dataframe like this:

name      link
apple    example1.com/dsa/es?id=2812168&width=1200/web/map&resize.html
banana.  example2.com/es?id=28132908&width=1220/web/map_resize.html
orange.  example3.com/es?id=3209908&width=1120/web&map_resize.html

Each name's ID is buried in the link, which may have different structure. However, I know that the pattern is 'id=' + 'what I want' + '&'

I wonder, is there a way to extract the id from link and put it back to the dataframe to get the following:

name      link
apple    2812168
banana.  28132908
orange.  3209908

I try to use this:

df['name'] = df['name'].str.extract(r'id=\s*([^\.]*)\s*\\&', expand=False)

but it returns a column with all nan

Also, there may be more than one & in the link

Answer 1

We can make use of positive lookbehind and positive lookahead :

df['link'] = df['link'].str.extract('(?<=id\=)(.*?)(?=\&)')


      name      link
0    apple   2812168
1  banana.  28132908
2  orange.   3209908

Details :

• (?<=id\=) : positive lookbehind on id=
• (.*) : everything
• (?=\&width) : positive lookahead on &width

Answer 2

I think Ids are always numbers, so this is somewhat cleaner:

df["link"] = df['link'].str.extract(r'id=(\d+)&', expand=False)
print(df)
#     name      link
#0   apple   2812168
#1  banana  28132908
#2  orange   3209908

Answer 3

Let tri split

df['link'].str.split('id=').str[1].str.split('&').str[0]
0     2812168
1    28132908
2     3209908
Name: link, dtype: object