CODEWITHSUNDEEP
perlawksed
user1987607
user1987607

Reputation: 2157

remove spaces, but keep tabs

I have a tab-delimited text file with 3 columns.

In some of the columns there are single or multiple spaces that I want to remove. But I want to keep the tab separation between each column and also the newline character.

I tried 

perl -lape 's/\s+//sg

but that removes all whitespaces, also the tab spaces

My file looks like this

col1 col2    col3
1    test    test
2       test test
3     test    test

And I want

col1   col2   col3
1      test   test
2      test   test
3      test   test

So I only want to keep the tabspaces between the different columns, but not the single spaces. I hope this is clear.

Upvotes: 3

Views: 2236

Answers (4)

glenn jackman
glenn jackman

Reputation: 246744

With awk, to reformat the output to use specifically a tab character,

awk -v OFS='\t' '{$1=$1}1' file

The odd-looking $1=$1 forces awk to rewrite the current record using the Output Field Separator (tab)

Upvotes: 1

Sobrique
Sobrique

Reputation: 53478

If it's just spaces, you can use ' ' instead of \s.

E.g. 

s/ //g;

Of course, given you're doing lape and -a means 'autosplit on whitespace' you could just:

perl -ane 'print join ("\t", @F );'

Upvotes: 4

fedorqui
fedorqui

Reputation: 289505

Just remove spaces, not \s which also matches tabs:

sed 's/ \+//g' file

And if you want to remove these spaces just if they occur after a tab, say:

sed 's/\t */\t/g' file

From perldoc perlretut:

\s matches a whitespace character, the set [\ \t\r\n\f] and others

Upvotes: 7

Hunter McMillen
Hunter McMillen

Reputation: 61513

You can create you own character class that is the negation of all things that are not spaces and a tab, this character class represents all characters that are whitespace - tabs:

perl -lape 's/[^\S\t]+//sg'

[ ... ] defines a character class

^ inside of [ ... ] negates this character class

\S represents everything not in \s

\t represents a tab character

Upvotes: 2

Related Questions