How to replace a pattern in a string?

Question

Hi I am trying to replace all the expressions containing 'www...' and 'http://..' with just 'URL'. I tried this but I am getting this error.

TypeError: expected string or buffer

My code is:

df['text_1'] = re.sub('((www\.[^\s]+)|(https?://[^\s]+))','URL',df['text'])

df[text] contains tweets, so I want to keep only the text in there. I am in Python 2 Thanks.

Is df[text] a list of tweets, i.e. a list of strings, or a single string? Have you tried ... = [re.sub('<regex>','URL',s) for s in df['text']]? — tobias_k, Commented May 15, 2017 at 22:05
in each value of df[text] there is one tweet. This is what are you asking? — Horacio Nesman, Commented May 15, 2017 at 22:09
Please clarify what data type df actually is. We know it's not a string and not a buffer, I'm assuming it's a pandas DataFrame. — acidtobi, Commented May 15, 2017 at 22:21
yes it is DataFrame, would you recommend me where can I read more about these differences? String, DataFrames and Buffer? Now I am a bit confuse about this. Thanks — Horacio Nesman, Commented May 15, 2017 at 22:33

acidtobi · Accepted Answer · 2017-05-15 22:14:10Z

2

Assuming df is a pandas DataFrame, don't use re.sub. Use pandas.DataFrame.replace instead:

df['text_1'] = df['text'].replace('((www\.[^\s]+)|(https?://[^\s]+))',
                                  'URL',
                                  regex=True)

This will generate a new column text_1 with all values of text replaced according to your regular expression.

answered May 15, 2017 at 22:14

acidtobi

1,3659 silver badges13 bronze badges

Add a comment |

Ryan Baker · Accepted Answer · 2017-05-15 22:08:42Z

1

It sounds like you're getting that error because you're not supplying a string or buffer as the third argument to re.sub.

>>> re.sub('\W', 'REPLACED', 'this is my text')
'thisREPLACEDisREPLACEDmyREPLACEDtext'
>>> re.sub('\W', 'REPLACED', None)
Traceback (most recent call last):
...
TypeError: expected string or buffer

Ensure that df['text'] contains a proper string before you try using it for re.sub

answered May 15, 2017 at 22:08

Ryan Baker

3031 silver badge8 bronze badges

That worked, yes I am using dataframes, thanks everybody
– Horacio Nesman
Commented May 15, 2017 at 22:17

Add a comment |

Collectives™ on Stack Overflow

How to replace a pattern in a string?

2 Answers 2

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Related