This is my pandas data-frame. I want to modify values of first column (System) by extracting just a,b,c,d. How can this be done in python
System mem
/vol/a/ 10
/vol/b/ 20
/vol/c/ 30
/vol/d/ 40
-
Are they always the last characters before /?Vaishali– Vaishali2017-11-14 04:37:25 +00:00Commented Nov 14, 2017 at 4:37
-
yeah.. there can be strings also in place of a,b,c,dria– ria2017-11-14 04:39:29 +00:00Commented Nov 14, 2017 at 4:39
-
as in multiple character? say 'abc'?Vaishali– Vaishali2017-11-14 04:41:30 +00:00Commented Nov 14, 2017 at 4:41
Add a comment
|
4 Answers
You can use .str.extract:
In [11]: df.System.str.extract("/vol/(.*?)/", expand=False)
Out[11]:
0 a
1 b
2 c
3 d
Name: System, dtype: object
4 Comments
ria
what if there is also sequences like /vol/abc/.aggr and we have to extract abc.
Andy Hayden
@madhuri make the regex pattern
"/vol/(.*?)/.*"Andy Hayden
Now I think about it the
? is not required here.
Jan
@AndyHayden: It is for strings like
/vol/a/.hidden/.git, see OP's comment above. A bit unclear what he really wants.Can be done in multiple ways, here is one
df['System'] = df['System'].str.split('/').str[-2]
System mem
0 a 10
1 b 20
2 c 30
3 d 40
Option 2:
df['System'] = df.System.str.replace('[/vol/|/]', '')
Andy Hayden already covered str.extract
3 Comments
ria
what if there is also sequences like /vol/abc/.aggr and we have to extract abc
Vaishali
str.extract would work best in that case, df['System'].str.extract('/vol/([A-Za-z]+)/', expand = False)
Vaishali
@madhuri, consider upvoting the answers and accepting the most helpful answer. It helps in closing the question and works as an appreciation for people who take time out to answer
Using str.extract:
import pandas as pd
df = pd.DataFrame({'System': ['/vol/a/', '/vol/b/', '/vol/c/', '/vol/d/'], 'mem': [10, 20, 30, 40]})
df['new_column'] = df['System'].str.extract(r'([^/]+)/?$')
print(df)
This yields
System mem new_column
0 /vol/a/ 10 a
1 /vol/b/ 20 b
2 /vol/c/ 30 c
3 /vol/d/ 40 d
1 Comment
ria
what if there is also sequences like /vol/abc/.aggr and we have to extract abc
By using rsplit you can have as much stuff in front as you'd like.
df.assign(New=df.System.str.rsplit('/', 2).str[-2])
System mem New
0 /vol/a/ 10 a
1 /vol/b/ 20 b
2 /vol/c/ 30 c
3 /vol/d/ 40 d
Comments
