CODEWITHSUNDEEP
pythonstringpandasmatch
jeangelj
jeangelj

Reputation: 4498

python pandas partial string match

I created a dataframe df where I have a column with the following values: 

category
20150115_Holiday_HK_Misc
20150115_Holiday_SG_Misc
20140116_DE_ProductFocus
20140116_UK_ProductFocus

I want to create 3 new columns

category                  |           A              |  B  |       C     
20150115_Holiday_HK_Misc     20150115_Holiday_Misc     HK    Holiday_Misc 
20150115_Holiday_SG_Misc     20150115_Holiday_Misc     SG    Holiday_Misc
20140116_DE_ProductFocus     20140116_ProductFocus     DE    ProductFocus
20140116_UK_ProductFocus     20140116_ProductFocus     UK    ProductFocus

In column A, I want to take out "_HK" - I think I need to manually code this, but this is fine, I have the list of all country codes

In column B, it's that very country code

Column C, is column A without the date in the beginning

I am trying something like this, but not getting far. 

 df['B'] = np.where([df['category'].str.contains("HK")==True], 'HK', 'Not Specified')

Thank you

Upvotes: 1

Views: 1178

Answers (2)

MaxU - stand with Ukraine
MaxU - stand with Ukraine

Reputation: 210832

you can use Series.str.extract() method:

# remove two characters (Country Code) surrounded by '_'
df['A'] = df.category.str.replace(r'_\w{2}_', '_')
# extract two characters (Country Code) surrounded by '_' 
df['B'] = df.category.str.extract(r'_(\w{2})_', expand=False)
df['C'] = df.A.str.extract(r'\d+_(.*)', expand=False)

Result:

In [148]: df
Out[148]:
                   category                      A   B             C
0  20150115_Holiday_HK_Misc  20150115_Holiday_Misc  HK  Holiday_Misc
1  20150115_Holiday_SG_Misc  20150115_Holiday_Misc  SG  Holiday_Misc
2  20140116_DE_ProductFocus  20140116_ProductFocus  DE  ProductFocus
3  20140116_UK_ProductFocus  20140116_ProductFocus  UK  ProductFocus

Upvotes: 5

Yuval Atzmon
Yuval Atzmon

Reputation: 5945

You can also use regex and apply

import re
df['A'] = df.category.apply(lambda x:re.sub(r'(.*)_(\w\w)_(.*)', r'\1_\3', x))
df['B'] = df.category.apply(lambda x:re.sub(r'(.*)_(\w\w)_(.*)', r'\2', x))
df['C'] = df.A.apply(lambda x:re.sub(r'(\d+)_(.*)', r'\2', x))

Result

                   category                      A   B             C
0  20150115_Holiday_HK_Misc  20150115_Holiday_Misc  HK  Holiday_Misc
1  20150115_Holiday_SG_Misc  20150115_Holiday_Misc  SG  Holiday_Misc
2  20140116_DE_ProductFocus  20140116_ProductFocus  DE  ProductFocus
3  20140116_UK_ProductFocus  20140116_ProductFocus  UK  ProductFocus

Upvotes: 1

Related Questions