Removing handles and links from string

[kiton]

Hello! My "Python quest" continues and I am seeking your help with the following task.

So, I have a data frame of 3199 rows with a string -text- column, e.g.:

df['text'].iloc[2001]

Output:
I miss biz!#ThingsYouDontWannaHearBeforeAKiss "Hope you like bacon and corn"@martensmi Love it. Jack and I r watching as wellYour compliments mean more than anyone else\'s.@DaveAbberger @markholston prince NEEDED to go... .196 in the postseason with like 3RBI\'s?!?! So glad he\'s gone... Now we need Inge back :)On this damn damnI\'m at Jimmys\' Lounge (Hazen, ND) http://t.co/ciYHl8J3x2I don\'t wanna fall in love to think about the future. I wanna fall in love and have fun because I\'m young and I deserve fun ya knowMy 2nd home? (@ Krause\'s SuperValu) http://t.co/4atauDdqfqVery blessed to say I have my first varsity start under my belt. Thanks to everyone who has supported me throughout the years.Pool pilates! (@ Hazen Swimming Pool) [pic]: http://t.co/zY9DIU6HHo"There are two tragedies in life, one is to lose your hearts desire, the other is to gain it." -George Bernard ShawI jus want this week to be over....@JeffGordonWeb WINS AT @MartinsvilleSwy YUSSS!!!!! #NASCAR

What code should I use in order to delete (1) Twitter handles -- i.e., words that start with "@", and (2) links that start with "http://"? 

I tried split and regex, but failed to run it successfully. Thank you in advance for help.

[zivoni]

I guess that you should use regex matching both links and @username to remove them.

I tried it with simple

pattern = r'(http://[^"\s]+)|(@\w+)'

where http://[^"\s]+ matches string starting with http:// and ending before following whitespace or ", @\w+ matches @ followed by "word" characters.

You can replace matches of regex in dataframe column "text" with

df.text.str.replace(pattern, "")

when i tried it on your text, result was

Output:
I miss biz!#ThingsYouDontWannaHearBeforeAKiss "Hope you like bacon and corn" Love it. Jack and I r watching as wellYour compliments mean more than anyone else\'s.  prince NEEDED to go... .196 in the postseason with like 3RBI\'s?!?! So glad he\'s gone... Now we need Inge back :)On this damn damnI\'m at Jimmys\' Lounge (Hazen, ND)  don\'t wanna fall in love to think about the future. I wanna fall in love and have fun because I\'m young and I deserve fun ya knowMy 2nd home? (@ Krause\'s SuperValu)  blessed to say I have my first varsity start under my belt. Thanks to everyone who has supported me throughout the years.Pool pilates! (@ Hazen Swimming Pool) [pic]: "There are two tragedies in life, one is to lose your hearts desire, the other is to gain it." -George Bernard ShawI jus want this week to be over.... WINS AT  YUSSS!!!!! #NASCAR'

Problem with such simple regex for twitter username is that it would catch parts of emails too - so it would be better to replace @\w+ with smarter pattern preserving emails

(?<=^|(?<=[^a-zA-Z0-9-_\.]))@([A-Za-z]+[A-Za-z0-9]+)

that i found on stackoverflow.

[kiton]

Thank you for response, zivoni. Your suggestion has worked fine (both with and without "adjustment" for emails). I just have one follow-up question (likely a noob one), why the resulting output is not a table (no columns with headings), e.g.:

Output:
0        he can't bc of meI'm at McQueen Village Apts ...
1       Just finished. Such a good book! Suggestions f...
2       We've been through the worse made it through t...
3       Something's ion Understand . have enough for m...
4       You didn't eat breakfast? Good you're grounded...
5       We going to see Carrie on the 19thI'm about to...