Problem with AWK

[linuxMaestro]

# awk '/([a-Z][0-9])*\@([a-Z][0-9])\.[a-z][a-z][a-z]?/' filename
awk: cmd. line:1: fatal: Invalid range end: /[a-Z]*\@[a-Z]\.[a-z][a-z][a-z]?/

I am trying to get email addresses from file filename, what am I doing wrong?

I have
([a-Z][0-9])* Match any number of letters or numbers
\@ Before an @
([a-Z][0-9])* then Match any number of letters or numbers
\. Before a .
[a-z][a-z][a-z]? then match either 2 or 3 characters

Why won't it work?

 

[chipperMDW]

> Invalid range end

I'd look at those [tt][a-Z][/tt] ranges and their invalid ends.

The end is supposed to come after the beginning. In ASCII, the capital letters come before the lowercase letters. So capital [tt]Z[/tt] comes before lowercase [tt]A[/tt].

However, [tt][A-z][/tt] is not right, either. Several characters ([tt][[/tt], [tt]\[/tt], [tt]][/tt], [tt]^[/tt], [tt]_[/tt], and [tt]`[/tt]) come between [tt]Z[/tt] and [tt]a[/tt]. You should specify them as separate ranges: [tt][a-z][/tt] and [tt][A-Z][/tt].

Even then, I still think there's something wrong with your idea of a valid e-mail address... I think there can be multiple dots in the part before the [tt]@[/tt], for one thing... I'd look at an RFC decribing valid e-mail addresses if I were you and I wanted this to work on any a-mail address.

 

[futurelet]

Try this one:

# Try to match email addresses.
BEGIN { a="[a-zA-Z0-9]"; A="[a-zA-Z]"
e_pat = "(^|[ \t])("a"+[.])*"a"+@"A A A"?([^a-zA-Z]|$)"
}
$0 ~ e_pat