How to remove after second period in a string using sed

Question

In my script, have a possible version number: 15.03.2 set to variable $STRING. These numbers always change. I want to strip it down to: 15.03 (or whatever it will be next time).

How do I remove everything after the second . using sed?

Something like:

$(echo "$STRING" | sed "s/\.^$\.//")

(I don't know what ^, $ and others do, but they look related, so I just guessed.)

^ is the beginning and $ is the end of an input/line. gnu.org/software/sed/manual/html_node/Regular-Expressions.html — matt
– matt, Commented Sep 22, 2018 at 9:18

ctac_ · Accepted Answer · 2018-09-22 14:31:07Z

5

I think the better tool here is cut

echo '15.03.2' | cut -d . -f -2

answered Sep 22, 2018 at 14:31

ctac_

2,5012 gold badges10 silver badges18 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Ed Morton Over a year ago

Yes it is... :-).

Jesse Over a year ago

Yes, a better tool and thanks. But, my question was about sed specifically, just sayin.

potong · Accepted Answer · 2018-09-22 18:00:04Z

1

This might work for you (GNU sed):

sed 's/\.[^.]*//2g' file

Remove the second or more occurrence of a period followed by zero or non-period character(s).

answered Sep 22, 2018 at 18:00

potong

59.3k6 gold badges55 silver badges92 bronze badges

4 Comments

potong Over a year ago

@EdMorton see here for details of why the flags number and g in the substitution command act as they do for GNU sed.

Ed Morton Over a year ago

I'm very familiar with those flags, I just don't understand why that particular combination is necessary given that regexp. For example, given input of a.b.c.d and the regexp \.[^.]* the 2nd . is between b and c so the 2 should match the leftmost longest matching substring which is .c.d so why then do you also need a g after it?

potong Over a year ago

@EdMorton using your example the 2nd matches .c the g flag then matches any following such regexs, in your case .d too. So the combination of number and g flag, matches from number to the end of the string.

Ed Morton Over a year ago

Ah, right, I was getting confused about what the [^.]* would match, thanks.

Ed Morton · Accepted Answer · 2018-09-22 13:08:25Z

$ echo '15.03.2' | sed 's/\([^.]*\.[^.]*\)\..*/\1/' 15.03

More generally to skip N periods:

$ echo '15.03.2.3.4.5' | sed -E 's/(([^.]*\.){2}[^.]*)\..*/\1/' 15.03.2 $ echo '15.03.2.3.4.5' | sed -E 's/(([^.]*\.){3}[^.]*)\..*/\1/' 15.03.2.3 $ echo '15.03.2.3.4.5' | sed -E 's/(([^.]*\.){4}[^.]*)\..*/\1/' 15.03.2.3.4

Collectives™ on Stack Overflow

How to remove after second period in a string using sed

3 Answers 3

2 Comments

4 Comments

1 Comment

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

2 Comments

4 Comments

1 Comment

Related