23

I have some text files which contain some columns separated by a various number of spaces, but instead I need one single tab as a separator. Is it possible to do in Bash?

  • Thanks for the great input, but i have some single spaces inside a column, so i have to avoid tabbing a single space. sorry for that ,isinformation. – user_unknown Feb 2 '11 at 22:49
27

To convert sequences of more than one space to a tab, but leave individual spaces alone:

sed 's/ \+ /\t/g' inputfile > outputfile

To do this for a number of files:

for inputfile in *
do
    sed 's/ \+ /\t/g' "$inputfile" > tmpfile && mv tmpfile "$inputfile"
done

or

for inputfile in *
do
    sed -i.bak 's/ \+ /\t/g' "$inputfile"
done

or

find . -type f -exec sed -i.bak 's/ \+ /\t/g' {} \;
  • sed: -e expression #1, char 1: unknown command: `.' – Aaron Franke Feb 10 at 0:40
  • @AaronFranke: What command did you try? None of the examples in my answer should produce that error. – Dennis Williamson Feb 10 at 1:25
  • Sorry, I should've clarified. The find one on the bottom. – Aaron Franke Feb 10 at 9:06
  • @AaronFranke: GNU sed doesn't like to have a space before the backup extension. I've edited my answer. Thanks for the report. – Dennis Williamson Feb 10 at 13:22
6

If your character is multiple tabs you can also use tr -s:

-s, --squeeze-repeats   replace each input sequence of a repeated character
                        that is listed in SET1 with a single occurrence

For example:

my_file.txt | tr -s " "

All white spaces will become one.

5

You can use sed to replace a number of spaces with a tab.:

Example to replace one-or-more-spaces with one tab:

cat spaced-file | sed 's/ \+/\t/g' > tabbed-file
  • The OP said the number of spaces was variable, so I don't think this solution will work. – Mikel Feb 2 '11 at 22:35
  • @Mikel. Oops. Thanks for pointing that out. I've edit the post to allow matching for variable spaces. – IvanGoneKrazy Feb 2 '11 at 22:45
  • Most useful answer here. – Luís de Sousa Jan 15 '16 at 14:07
3

The easiest answer using only bash is:

while read -r col1 col2 col3 ...; do
    echo -e "$col1\t$col2\t$col3..."
done <file

If there are a variable number of columns, you can do this, but it will only work in bash, not sh:

while read -r -a cols; do
    (
        IFS=$'\t'
        echo "${cols[*]}"
    )
done <file

e.g.

while read -r -a cols; do
    (
        IFS=$'\t'
        echo "${cols[*]}"
    )
done <<EOF
a b   c
d   e    f
  g h i
EOF

produces:

a   b   c
d   e   f
g   h   i

(there is a tab in between each, but it's hard to see when I paste it here)

You could also do it using sed or tr, but notice that the handling of blanks at the start produces different results.

sed:

$ sed 's/  */\t/g' << EOF
a b   c
d   e    f
  g h i
EOF
a       b       c
d       e       f
        g       h       i

tr:

$ tr -s ' ' '\t' <<EOF
a b   c
d   e    f
  g h i
EOF
a       b       c
d       e       f
        g       h       i
2

perl -p -i -e 's/\s+/\t/g' *.txt

2

Try the following SED script:

 sed 's/  */<TAB>/g' <spaces-file > tabs-file

Where <TAB> is pressing the TAB key.

0

This is a very simple solution:

    sed -E 's/\s+/\t/g' your_file > new_file

sed basically works in this manner (sed 's/old_pattern/new_pattern/g'). In this case the old pattern is "\s+" which means find space "s" one or more time "+" and the back slash "\" to interpret that as regular expression.
The new pattern is tab "\t" which is written in regular expression format and the "g" is apply the replacement to all lines "globally".

  • 1
    Hello and welcome to superuser. You should take the time to explain your solution. For someone not familiar with *nix systems, sed and regular expressions, this looks like a pile of weird characters. – Mogget Apr 3 '16 at 21:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.