Pandas regex to return any string that contains U or UN with a digit

user_6353522 0
Viewed 43

I'm trying to create a new column with values from another columns string What I want is to create a new column with unit values.

The position of the units can vary.

The examples of my strings are

this is a string and we have 4U to use
this is another string 5UN
only 6U to use today

I need to extract the numbers that are joined to both U and UN since the positions vary.

df['test_units'] = df['ITEM_DESC'].str.get(r'\(*U.*?\)',)
df['test_units']

This is my regex but I return only nan values.

How do I just return the number that's joined to a U Or UN?

pandas django python regex
2 Answers

You may use

df['test_units'] = df['ITEM_DESC'].str.extract(r'\b(\d+)UN?\b')

See the regex demo. Note the unescaped pair of parentheses that form a capturing group whose value is returned by Series.str.extract.

The regex matches:

  • \b - a word boundary
  • (\d+) - Group 1: one or more digits
  • U - U
  • N? - an optional N
  • \b - word boundary

Pandas test:

import pandas as pd
cols={'ITEM_DESC': ['this is a string and we have 4U to use','this is another string 5UN','only 6U to use today']}
df = pd.DataFrame(cols)
df['test_units'] = df['ITEM_DESC'].str.extract(r'\b(\d+)UN?\b')

Output:

>>> df
                                ITEM_DESC test_units
0  this is a string and we have 4U to use  4        
1  this is another string 5UN              5        
2  only 6U to use today                    6        
>>>
user_38329700

Why do you want U OR UN? This is because you might have something like 5UD that you don't want? In any case:

patt = re.compile(r"\d+UN?\b")
user_108099680
Related