i need to strip the url from a given string. The only solution i have is:
_url="http://download.enlightenment.org/rel/apps/econnman/econnman-1.1.tar.gz"
_l=${_url%.*/*} # http://download.enlightenment
_l=${#_l} # 29
_url=${_url:0:${l}+4} # http://download.enlightenment.org
But this will fail on every none 3 chars length TLD and it's just the wrong way.
thx a lot.
You can use grep:
$ echo "$_url" | grep -Eo '^http[s]?://[^/]+'
http://download.enlightenment.org
You can use expr with a regex:
$ echo `expr "$_url" : '\(http://[^/]*\)'`
http://download.enlightenment.org
Or, use awk:
echo "$_url" | awk -F/ 'BEGIN{OFS=FS} {print $1 OFS OFS $3}'
http://download.enlightenment.org
You can use cut:
echo "$_url" | cut -d/ -f1-3
http://download.enlightenment.org
cut is probably the easiest to get the rest of the url as well:
$ echo "$_url" | cut -d/ -f4-
rel/apps/econnman/econnman-1.1.tar.gz
Or, completely internal to Bash:
$ [[ $_url =~ ^([^:]+://[^/]+)/?(.*)$ ]] && server="${BASH_REMATCH[1]}"
$ echo "$server"
http://download.enlightenment.org
and "${BASH_REMATCH[2]}" has the rest of the url.
To extract a substring from a value already contained in a shell variable, use Bash's regex-matching operator, =~, which supports extended regular expressions:
Note: dawg's answer contains solutions that are better suited to input from a file or stdin with multiple inputs.
They incur startup cost due to involving child processes, but for collections of inputs that is well worth it, because external utilities are much more efficient at processing larger input sets.
_url='http://download.enlightenment.org/rel/apps/econnman/econnman-1.1.tar.gz'
[[ $_url =~ ^https?://[^/]+ ]] && _url="${BASH_REMATCH[0]}"
echo "$_url" # -> 'http://download.enlightenment.org'
^https?://[^/]+ ]] matches any string that starts with (^) literal http:// or https:// and matches the longest nonempty (+) run of characters that follows that doesn't include / ([^/]+).
Built-in array BASH_REMATCH contains the results of the most recent application of the =~ operator, with the first element (with index 0) containing whatever the regex matched as a whole.
(Subsequent elements would contain what parenthesized
sub-expressions (a.k.a capture groups) matched, but in this case we're not using any).
I don't know if this works for bash, but it works for pcre regex engine.
(?<=:\/\/)(.*)(?=\/)
Finds all text between first / and second /. Works for https://google.com/ but doesn't for google.com/ or https://google.com. Depends on what you need.
