I have string such as
username/ticket-12345/feature
and i want to extract just
ticket-12345
from bash. the forrmat of this string could be anything.... e.g.
'my string ticket-12345'
and 'ticket' could be a mixture of lower case and upper case. Is this possible to do from bash? I've tried searching for this particular case but i can't seem to find an answer...
Here is a pure bash regex method:
re='[[:alpha:]]+-[0-9]+'
s='username/ticket-12345/feature'
[[ $s =~ $re ]] && echo "${BASH_REMATCH[0]}"
ticket-12345
s='my string ticket-12345'
[[ $s =~ $re ]] && echo "${BASH_REMATCH[0]}"
ticket-12345
ticket can be any alpha string, or can only consist of the letters T/t, I/i, C/c, K/k, and so forth[A-Za-z] to [[:alpha:]], since the A-Z approach is collation-order/locale-dependent.
ticket in mixed cases.
The shell's built-in ERE (extended regular expression) support is adequate to the task:
ticket_re='[Tt][Ii][Cc][Kk][Ee][Tt]-[[:digit:]]+'
string='my string ticket-12345'
[[ $string =~ $ticket_re ]] && echo "Found ticket: ${BASH_REMATCH[0]}"
nocasematch with ticket-[[:digit:]]+ to ignore the case of "ticket", although that becomes awkward if you need to limit its scope to just this one match.With the -o flag grep and its friends display only the found matches. You can use
egrep -io 'ticket-[0-9]+' file.txt
to find the tickets from your input text.
-o being a non-POSIX extension), as-updated -- whether this is more appropriate than the pure bash method depends on usage context (grep can scan through a large file much more quickly, but if each call is evaluating only a single line, then it'll be far slower than the shell-builtin alternative on account of the overhead required to fork and exec() an external process).