-1

I have a text file (approx. 300'000 lines) with a unique column exhibiting 2 to 6 fields separated by a comma.

A,BB,CC  
EF,GHI,DKLM,OP          
Q,RS,TUV 
W,XY

What I would I like to achieve:

A,B,C  
EF,HI,KLM,P         
Q,S,UV 
W,Y

Said differently, I would like to remove in each line, the FIRST character after each comma.

Thanks in advance for your time and help. Best regards,

Laurent

Opus_francigenum
  • 31
  • 5

1 Answers1

3
$ cat file
A,BB,CC
EF,GHI,DKLM,OP
Q,RS,TUV
W,XY

$ sed 's/,./,/g' file
A,B,C
EF,HI,KLM,P
Q,S,UV
W,Y

The regular expression ,. matches a literal comma followed by any other character whatsoever. The sed expression s/,./,/g substitutes each comma and the following character with just a single comma, effectively removing that other character. The substitution is carried out for each non-overlapping match on every line.

Redirect the output to a new file, or use sed with -i in an appropriate way (see How can I achieve portability with sed -i (in-place editing)?).

Kusalananda
  • 320,670
  • 36
  • 633
  • 936
  • Dear Kusalananda. Thanks for your response, I appreciate that you took some of your time to help me. I have a question following your post. What do you mean by "The substitution is carried out for each non-overlapping match on every line"? If in a same line, I have twice the same substitution, for instance [,X] and later on the same line [,X], it will not work? – Opus_francigenum Jul 19 '21 at 10:15
  • 1
    @Opus_francigenum In your example, there is never any problem since all occurrences of `,X` are non-overlapping. If you had `,,X`, then the second comma would be removed, not the `X`. This is because in `,,X`, the two substrings `,,` and `,X` are _overlapping_. If you have cases like this and want to avoid removing commas, then use `s/,[^,]/,/g` instead as the `sed` expression. This changes the meaning from "change any comma and the following character with a comma" to "replace any comma and the following non-comma with a comma". – Kusalananda Jul 19 '21 at 10:22
  • Crystal clear. Many thanks Kusalananda for your time ... and patience :-) Best regards to you. – Opus_francigenum Jul 19 '21 at 10:35
  • @Opus_francigenum Good! If this solves your issue, please consider [accepting the answer](https://unix.stackexchange.com/help/someone-answers). Accepting an answer marks the issue as resolved. – Kusalananda Jul 19 '21 at 10:38