[techtalk] need help with a shell script

[Andrew Wendt]

On Tue, 26 Sep 2000, Laurel Fan wrote:

>Here's a way to get the [domain-minus-top-level-domain] part:
>
>   echo $DOMAIN | sed -e 's/\.[^\.]*$//'

If you're using bash, you can also do this directly from the shell:

echo "${DOMAIN%.*}"

This will take the shortest amount of text matching the glob pattern .* from 
the end of your variable and throw it away.

You can also use %% instead of % to throw away the longest amount of text 
matching the pattern. Similarly, # and ## throw away the shortest and longest
match from the beginning of your variable.

So:
---snip---
$ DOMAIN=foo.bar.com
$ echo "${DOMAIN%.*}"
foo.bar
$ echo "${DOMAIN%%.*}"
foo
$ echo "${DOMAIN#*.}"
bar.com
$ echo "${DOMAIN##*.}"
com
---snip---

>now for the [first-letter-of-domain] part:
>
>  echo $DOMAIN | cut -c 1

You can also do this from the shell.

---snip---
$ echo "${DOMAIN:0:1}"
f 
---snip---

The first number is which letter to start at (counted from zero), and the 
next is the length you want.

By the way, in your usage of sed here:
echo $DOMAIN | sed -e 's/\.[^\.]*$//'

You needn't escape the dot inside your square brackets. And in fact by trying 
to I think you have made a pattern which matches a dot followed by some text 
which isn't another dot or a backslash. So it wouldn't properly work with a 
name like "foo.bar.c\m".

Which shouldn't matter since you don't see many backslashes in domain names. 
:-)

TTFN
Andy