Hi guys

i have a file with the following content

$ cat num2
100
101
1001
10001
100001
1000001

i'd like to match only 1001, 10001, 100001

i did this but it doesn't work

$ grep 10\{2,3\}1 num2
$

any idea?

it worked with egrep

$ egrep 10\{2,4\}1 num2
1001
10001
100001

it may be more precise to use something like this:

1) grep and sed support two levels of regular expressions "basic", and "extended". Only the extended regex offers things like "{}". To enable them use grep -E and sed -r. egrep is a synonym for grep -E and is apparently deprecated.

However, when in basic regex mode, gnu grep and sed also allow you to activate extended regex characters individually by preceding them with backslashes. Escaping them when in extended regex mode, OTOH, does the reverse, and disables them. The grep man page has a whole section on basic vs. extended regex.

So your expression is actually ok in this regard (although it really needs to be anchored on each end, as rosehosting posted).


2) The real problem you are encountering is actually the behavior of the shell. The shell reduces all backslashed characters to their literal equivalents, so "\{" is converted back to "{" before the command is executed. You're right back to using basic regex. The backslash also needs to be protected so that it can be properly passed to grep.

This is why you should always quote the expression in commands like this.