Extract pattern from a string in a bash script

Question

Somewhat new to bash.

I've experimented with parameter expansions, grep, sed and echo to solve this problem, but cannot quite work it out.

I'm trying to extract a certain pattern from $PWD in a bash script.

Let's say there could be a variety of full paths:

/home/files/tmp8
/home/tmp28/essential
/home/tmp2/essential/log
/home/files/tmp10/executables
/tmp8/files/whatever/etc

In every instance, I want to extract any string that contains "tmp" followed by 1 or more integers.

So, in each instance in which $PWD is processed, it will return "tmp8", "tmp28", "tmp2" etc.

Explanations for how the functions/operators work in regards to solving this issue would also be greatly appreciated.

lojoe · Accepted Answer · 2017-01-04 18:37:43Z

6

You can use regular expressions in bash to extract a pattern from any path string. See the following example:

if [[ "$PWD" =~ ^.*(tmp[0-9]+).*$ ]]
then
    printf "match: ${BASH_REMATCH[1]}\n"
else
    printf "no match: $PWD\n"
fi

The regular expression defines a group in the round parenthesis. If the expression matches the matching group (tmp with at least one digit following) it will be stored by Bash in the array BASH_REMATCH.

edited Jan 4, 2017 at 18:37

answered Jan 4, 2017 at 18:28

lojoe

5311 gold badge3 silver badges14 bronze badges

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

Comments

anubhava · Accepted Answer · 2017-01-04 18:22:21Z

1

Use grep -o to show only matched text using regex tmp[0-9]* i.e. literal text tmp followed by 0 or more digits:

grep -o 'tmp[0-9]*' file
tmp8
tmp28
tmp2
tmp10
tmp8

answered Jan 4, 2017 at 18:22

anubhava

790k67 gold badges603 silver badges671 bronze badges

6 Comments

anubhava Over a year ago

Yes sure you can do str=$(grep -o 'tmp[0-9]*' <<< "$PWD") to assign extracted pattern to a variable str

anubhava Over a year ago

grep -o 'tmp[0-9]*' <<< "$PWD" only returns matched string i.e. tmp<integer> not the full input.

lojoe Over a year ago

The regex 'tmp[0-9]*' is not correct. That one also matches 'tmp' without a following digit. The star means any cardinality including zero.

anubhava Over a year ago

Of course I know that. If OP always wants 1 or more digits after tmp then OP can use; grep -Eo 'tmp[0-9]+' <<< "$PWD"

Gabriel Saul Over a year ago

@anubhava Yes, it works with "pset" followed by one or more numeric digits. Thank you.

|

Collectives™ on Stack Overflow

Extract pattern from a string in a bash script

2 Answers 2

Comments

6 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

6 Comments

Your Answer

Sign up or log in

Post as a guest

Related