Bash scripting, regex in if statement

Question

I'm pretty new to bash scripting and regexp and have a question. I want to check to see if my variable $name starts with a-d, e-h, i-l etc and do some stuff accordingly. If the string starts with "the." or "The." it should check the first letter after the period.

My problem is that if $name consists of "the.anchor" both the a-d0-9 and q-t will be true. Do you guys have any idea what's wrong?

if [[ $name =~ ^([tT]he\.)?[a-dA-D0-9]+ ]]; then
    do some stuff
fi

if [[ $name =~ ^([tT]he\.)?[e-hE-H]+ ]]; then
    do some stuff
fi

if [[ $name =~ ^([tT]he\.)?[i-lI-L]+ ]]; then
    do some stuff
fi

if [[ $name =~ ^([tT]he\.)?[m-pM-P]+ ]]; then
    do some stuff
fi

if [[ $name =~ ^([tT]he\.)?[q-tQ-T]+ ]]; then
    do some stuff
fi

if [[ $name =~ ^([tT]he\.)?[u-wU-W]+ ]]; then
    do some stuff
fi

if [[ $name =~ ^([tT]he\.)?[x-zX-Z]+ ]]; then
    do some stuff
fi

Thanks in advance!

Boris the Spider · Accepted Answer · 2014-02-07 13:12:25Z

2

Your first part it optional:

([tT]he\.)?

So the.anchor matches the pattern ^([tT]he\.)?[a-dA-D0-9]+ because the the. matches `^([tT]he\.)? and the a matches [a-dA-D0-9]+. It matches ^([tT]he\.)?[q-tQ-T]+ because ^([tT]he\.)? is optional an t matches [q-tQ-T]+. Note not the whole input is consumed by the second pattern, in fact only the first character is grabbed.

You can verify this by having bash echo the match:

echo "${BASH_REMATCH[0]}"

Which should print the.anchor in the first case and t in the second.

You do not have an end anchor on the pattern so only part of the input needs to be matched. If you made the second pattern ^([tT]he\.)?[q-tQ-T]+$ then it would not match.

Alternatively you could make the the first part possessive - ^([tT]he\.)?+. This will mean that if the engine matches the first expression it will not be unmatched. In the latter case ^([tT]he\.)?+ will grab the the. and then not release it when [q-tQ-T]+ fails; this will cause the match to fail.

edited Feb 7, 2014 at 13:12

answered Feb 7, 2014 at 13:00

Boris the Spider

61.4k6 gold badges111 silver badges177 bronze badges

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

3 Comments

JonnyQuest Over a year ago

Thanks for answering. If I change the q-t to ^([tT]he\.)?[q-tQ-T]+$ it wont match if name is "the.tally"

Boris the Spider Over a year ago

@JonnyQuest no, it will not. You could also make the first part possessive so that if it is matched it will not be backtracked.

JonnyQuest Over a year ago

Found an "answer" to my question. Will accept my own answer in 2 days when I can. Thanks again!

JonnyQuest · Accepted Answer · 2014-02-07 13:21:42Z

0

I figured out a way to fix my problem by using elif statements and putting the q-t part as the last one

answered Feb 7, 2014 at 13:21

JonnyQuest

214 bronze badges

Comments

John B · Accepted Answer · 2014-02-07 14:13:43Z

0

I think the ? can be removed as the if statement is already doing the test. The + matches the preceding item at least once and would only be needed if you want to match more than one instance of the letters.

You can do it like this:

if [[ $name =~ ^[tT]he\.[a-dA-D0-9] ]]; then
    do some stuff
fi

The condition will only return true if the first character after ^[tT]he\. is [a-dA-D0-9].

However, I tend to think case is a cleaner solution than if statements when matching lists of characters against variables.

case $name in
    [tT]he\.[a-dA-D0-9]*)
        do some stuff
        ;;
esac

edited Feb 7, 2014 at 14:13

answered Feb 7, 2014 at 13:33

John B

3,6661 gold badge19 silver badges22 bronze badges

Collectives™ on Stack Overflow

Bash scripting, regex in if statement

3 Answers 3

3 Comments

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

3 Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related