check for string format in bash script

Question

I am attempting to check for proper formatting at the start of a string in a bash script.

The expected format is like the below where the string must always begin with "ABCDEFG-" (exact letters and order) and the numbers would vary but be at least 3 digits. Everything after the 3rd digit is a do not care. Expected start of string: "ABCDEFG-1234"

I am using the below code snippet.

[ $(echo "$str" | grep -E "ABCDEFG-[0-9][0-9][0-9]") ] && echo "yes"

str1 = "ABCDEFG-1234"

str2 = "ABCDEFG-1234 - Some more text"

When I use str1 in place of str everything works ok and yes is printed.

When I use str2 in place of str i get the below error

[: ABCDEFG-1234: unary operator expected

I am pretty new to working with bash scripts so any help would be appreciated.

The smallest change necessary to make your code work would be just to add quotes and a -n: [ -n "$(echo "$str" | grep -E "ABCDEFG-[0-9][0-9][0-9]")" ] && echo "yes" -- that way the output from grep isn't string-split and passed to [ ] as separate arguments, which was the immediate cause of your error. That said, that approach is unnecessarily inefficient and baroque, when grep can directly emit a true or false exit status based on whether it matched, with no need for [ ] or output parsing at all, and when the shell can do comparisons without needing any external tool like grep. — Charles Duffy
– Charles Duffy, Commented Oct 3, 2014 at 20:27
(...and the -n is only necessary in some pathological cases; in the common case, the quotes would be enough). — Charles Duffy
– Charles Duffy, Commented Oct 3, 2014 at 20:29

Charles Duffy · Accepted Answer · 2014-10-03 20:37:09Z

6

If this is bash, you have no reason to use grep for this at all; the shell has built-in regular expression support.

re="ABCDEFG-[0-9][0-9][0-9]"
[[ $str =~ $re ]] && echo "yes"

That said, you might want your regex to be anchored if you want a match in the beginning rather than anywhere in the content:

re="^ABCDEFG-[0-9][0-9][0-9]"
[[ $str =~ $re ]] && echo "yes"

That said, this doesn't need to be an ERE at all -- a glob-style pattern match would also be adequate:

if [[ $str = ABCDEFG-[0-9][0-9][0-9]* ]]; then echo "yes"; fi

edited Oct 3, 2014 at 20:37

answered Oct 3, 2014 at 20:22

Charles Duffy

299k43 gold badges441 silver badges497 bronze badges

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

Comments

iterexet · Accepted Answer · 2014-10-03 20:22:52Z

0

Try grep -E "ABCDEFG-[0-9][0-9][0-9].*"

answered Oct 3, 2014 at 20:22

iterexet

112 bronze badges

1 Comment

Charles Duffy Over a year ago

Doesn't solve the problem, which is due to failing to quote the output of grep into a single token for being passed to test. (Of course, if one used grep -q and ran if echo "$str" | grep -q -E ...; then..., then it'd be moot; involving test at all is overcomplicating things unnecessarily).

Collectives™ on Stack Overflow

check for string format in bash script

2 Answers 2

Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related