0

I have a lot of tab-separated data, and this means when I import that I'm getting a lot of columns (too many for Google Drive and OpenOffice).

I think what I need to do is find and replace all tabs with line breaks (could be wrong).

*EDIT - I ended up figuring this out by using a different method. Thanks everyone for your help. :)

Improve this question
1
  • Perhaps you can answer your own question, so that we all can benefit?
    – Jacob Jan Tuinstra
    Dec 10, 2012 at 6:26

3 Answers 3

Reset to default
3

On OS X you have to use $'' to insert \t or \n and escape \n in the replace pattern.

echo $'1\t2\n3\t4' | sed $'s/\t/\\\n/g'

But don't linefeeds already separate lines in TSV files? You could also use TextEdit to replace the tabs with some other characters like semicolons.

Improve this answer
2
  • Hi, need some clarification on this one. What would be the command to achieve this? Thanks for your help.
    – Drewdavid
    Nov 28, 2012 at 19:29
  • And yes, but spreadsheets are interpreting the tabs as columns, and there are way too many columns to import and so they are getting cut off.
    – Drewdavid
    Nov 28, 2012 at 19:31
0

You can try using a regular expression:

sed 's/\t/\n/' input.txt > output.txt

This will replace tabs \t with newlines \n.

Improve this answer
2
  • Hi and thank you. Is this something I am to do in Terminal? Thanks :)
    – Drewdavid
    Nov 25, 2012 at 19:40
  • This didn't work as expected. As below, except that only the first http was changed to hnnp, all the rest of the file seems unchanged. Thanks for your help.
    – Drewdavid
    Nov 26, 2012 at 17:59
0

To expand on root's response, you actually may need to add the /g at the end to do all the happenings.

You could also write it to the same file with an auto-backup flag.

sed -i.orig 's/\t/\n/g' file.txt

The ending /g will cause it to hit each happenstance instead of the first per line (if you need that).

The -i.orig writes the change to the same file after making a backup of the original (in this case file.txt.orig). That is just a matter of preference for me rather than having to then rename the original and the new one to take its place.

Update

I attached a screenshot to show it working for me... You're sure the file is tab delineated?

enter image description here

Improve this answer
6
  • Hi, thanks for your addition/recommendation. As above, is this something I am to perform in Terminal? Thanks will let you know how it goes. :)
    – Drewdavid
    Nov 25, 2012 at 19:43
  • Yes, this would be done in the Terminal. And the .orig would be located in the same directory as whatever input file you are working with.
    – nerdwaller
    Nov 25, 2012 at 19:45
  • Hi, this didn't exactly work as planned. It replaced all of the "t" characters in the text file with "n"s, so my URLs are "hnnp" for example. I copied and pasted the code, only changing the filename.
    – Drewdavid
    Nov 26, 2012 at 17:56
  • @Drewdavid Attached a screenshot to show it working. Perhaps the line codings aren't unix based, what was the file made using?
    – nerdwaller
    Nov 26, 2012 at 18:05
  • Thanks, I didn't clearly mention that I'm using Terminal on Mac OS X. The answer above is attempting to help but I'm not quite there yet.
    – Drewdavid
    Nov 28, 2012 at 19:29

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .