Regex extract substring after '=' sign

Question

I saw many examples, but for some reason it still does not work for me.

This is the command I'm executing:

NUMBER=$(docker logs vault | grep Token)
NUMBER=${NUMBER##*": "}
NUMBER=$(echo $NUMBER | sed 's/^token=(.*)$//g')
echo $NUMBER

I want to get the value after '=', which is a string basically. I tried using GREP, and other regex's but I either get nothing, or just the original string.

Please advise.

You're a) using `()` capture groups with Basic Regular Expressions, where they have to be `` (or use ERE instead, `sed -E`/`sed -r`) and b) you're removing the complete line, but you want to re-insert the captured group again. — Benjamin W., Feb 20 '18 at 20:03
Better to use `cut` here `echo 'token=dsa32e3' | cut -d= -f2` — anubhava, Feb 20 '18 at 20:04
Or, if supported, `number=$(echo token=dsa32e3);number=${number#*=}` — choroba, Feb 20 '18 at 20:10
in bash: `NUMBER=${tk/*=/} ; echo $NUMBER`, assuming tk="token=dsa32e3" — user unknown, Feb 20 '18 at 20:25
@anubhava, thank you! please post an answer so I'll accept it. — David Faizulaev, Feb 20 '18 at 20:25

Nisba · Answer 1 · 2018-02-20T20:22:50.150

1

With sed you can simply remove the token=, with

NUMBER=$(echo token=dsa32e3 | sed 's/^token=//g') 
echo $NUMBER

Other non-regexp based alternatives are possible, as other users pointed out.

Another fun possibility is using the negative lookbehind, not supported by sed, so I used perl.

NUMBER=$(echo token=dsa32e3 | perl -pe 's/.*(?<=token=)([a-z0-9]*)/$1/g')
echo $NUMBER

edited Feb 20 '18 at 20:22

answered Feb 20 '18 at 20:13

Nisba

2,838
1
21
45

score 1 · Accepted Answer · answered Feb 20 '18 at 20:28

1

To get text after a delimiter better to use cut instead of sed as in this example:

echo 'token=dsa32e3' | cut -d= -f2

dsa32e3

-d= sets delimiter as = for cut
-f1 makes cut print first field

answered Feb 20 '18 at 20:28

anubhava

713,503
59
514
593

I encounter another at the moment, I get back the string I wanted but with an additional ANSI character: "48372d06-03a5-f2a0-8428-ecd58cf57424\u001b[0m", what is this '\u001b[0m' string? and how do I get rid of it? – David Faizulaev Feb 21 '18 at 11:34
Is this `\u001b[0m` part of your original string as well? – anubhava Feb 21 '18 at 14:53
Nope, it's seems to be added after I use the grep command. – David Faizulaev Feb 21 '18 at 14:59
But there is no `grep` here.. Just a `cut` command. If your script has more than it then I suggest you provide a sample input and your script by editing your question so that I can investigate and suggest you a fix. – anubhava Feb 21 '18 at 15:01
1

I've added the additional command I perform. – David Faizulaev Feb 21 '18 at 15:25
ok can you tell me output of this command: `docker logs vault | awk -F= '/Token/{print $2}' | cat -A` – anubhava Feb 21 '18 at 15:27
Let us [continue this discussion in chat](http://chat.stackoverflow.com/rooms/165576/discussion-between-david-faiz-and-anubhava). – David Faizulaev Feb 21 '18 at 16:03
1

Solved it using sed -r "s/\x1B\[([0-9]{1,2}(;[0-9]{1,2})?)?[m|K]//g" The string at the end was ANSI colour code. – David Faizulaev Feb 22 '18 at 09:23
Oh ok glad you made it work. cheers – anubhava Feb 22 '18 at 09:42

Regex extract substring after '=' sign

2 Answers2