Regex Matching for Bash

Question

I have potential inputs that will come in from a read -e -p command in a bash script. For example, the user would type L50CA. Some other possibilites that the user could type in are: K117CB, K46CE2, or V9CE1.

I need to break up what was read in. I read in like this:

read -e -p "What first atom? " sel1

then I would like to make an array like this (but this will not separate):

arr1=($sel1)

But I need to separate the array so that

${arr1[0]} is equal to L ${arr1[1]} is equal to 50 and ${arr1[2]} is equal to CA

This separation has to work with the other possible user input formats like the ones listed above. Regex seems to be the way to do this. I can isolate the first two matches of the input with the following regular expressions: ^\D and \d*(?=\w)

I need help matching the third component and implementing it into an array. Alternatively, it is fine to to break up the user input into three new variables. Or we can place a space between each of the matches so L50CA is converted to L 50 CA because then arr1=($sel1) will work.

Thanks for your help.

What do K46CE2 split into? K, 46, CE2? Or K, 46, CE, 2? — Etan Reisner
– Etan Reisner, Commented Sep 21, 2014 at 22:41
It would split into K 46 CE2. It is fine if they come out separate like they did with jm66's answer because it is trivial to just combine the 3rd and 4th entry. — PhysicalChemist
– PhysicalChemist, Commented Sep 21, 2014 at 22:43

Etan Reisner · Accepted Answer · 2014-09-21 22:44:37Z

2

Bash only solution:

for sel in L50CA K117CB K46CE2 V9CE1; do
    [[ "$sel" =~ "^(\w)([0-9]+)(.*)" ]]
    printf '%s - ' "${BASH_REMATCH[@]}"
    printf \\n;
done

answered Sep 21, 2014 at 22:44

Etan Reisner

81.6k8 gold badges119 silver badges154 bronze badges

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

Comments

clt60 · Accepted Answer · 2014-09-21 22:28:37Z

1

The

for sel in L50CA K117CB K46CE2 V9CE1
do
        arr=($(sed 's/\([0-9][0-9]*\)/ \1 /g'<<<"$sel"))
        echo "${arr[@]}"
done

prints

L 50 CA
K 117 CB
K 46 CE 2
V 9 CE 1

answered Sep 21, 2014 at 22:28

clt60

64.3k17 gold badges114 silver badges206 bronze badges

1 Comment

PhysicalChemist Over a year ago

Awesome thanks so much. All I needed to do here is combine ${arr[2]}${arr[3]}. Great solution.

steffen · Accepted Answer · 2014-09-21 22:57:17Z

1

In bash using string manipulation:

 ~$ sel1=L50CA
 ~$ part1=$(expr match $sel1 "\([A-Z]\+\).*")
 ~$ part2=$(expr match $sel1 "[A-Z]*\([0-9]\+\).*")
 ~$ part3=$(expr match $sel1 "[A-Z]*[0-9]*\([A-Z]*\)")
 ~$ echo $part{1,2,3}
 L 50 CA
 ~$ arr=($part{1,2,3})
 ~$ echo ${arr[@]}
 L 50 CA

edited Sep 21, 2014 at 22:57

answered Sep 21, 2014 at 22:51

steffen

17.4k4 gold badges47 silver badges87 bronze badges

4 Comments

PhysicalChemist Over a year ago

I did not even know string manipulation existed in bash...how elegant. Thanks for the great solution.

Etan Reisner Over a year ago

That's not bash that's expr.

steffen Over a year ago

@EtanReisner en.wikipedia.org/wiki/Expr shows it's incorporated into the shell as a built-in command. (Here type expr gives expr is /usr/bin/expr though.)

Etan Reisner Over a year ago

Evaluating expressions is a builtin. expr is not. expr existed before the shell could do that sort of thing as I understand it (which would seem to be what wikipedia was trying to state there).

Collectives™ on Stack Overflow

Regex Matching for Bash

3 Answers 3

Comments

1 Comment

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

1 Comment

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related