Question about a RegEx

[thehemi]

Let's say I have a phrase, and I want to be able to catch
the phrase via a regular expression whenever it might have
a non-normal character inserted.

Let's say I want to catch anything from...

"M_ake Money" to "Make Mone_y"

Obviously w/ the _ (or similar) inserted anywhere in between.
Is there a way to write a regex that may catch the situation?

What I'm trying to accomplish is probably real obvious...

[vsemaska]

One way I can think is to use sed to remove those characters in your example:

SUBJ="M_ake Money"
tmp="`echo "$SUBJ" | sed "s/[_-]//g"`"
if [ "$tmp" = "Make Money" ]; then echo "Spam"; fi

This example removes underscores and dashes.

[thehemi]

Thanks for the suggestion.

But ideally I do not want "Make Money" to be a match.
I only want matches when someone is trying to be sly.

[Cabhan]

So let me get this straight:

You want to match any block of text that has an (for this instance) underscore in it?

/_/

alex@danu ~ $ echo "Make Money" | grep '_'
alex@danu ~ $ echo "M_ake Money" | grep '_'
M_ake Money

So to make this simpler, you can define the acceptable characters, and take the complement:

alex@danu ~ $ echo "Make Money" | grep '[^a-zA-Z ]'
alex@danu ~ $ echo "M_ake Money" | grep '[^a-zA-Z ]'
M_ake Money

Does this help?

[vsemaska]

Thanks for the suggestion.

But ideally I do not want "Make Money" to be a match.
I only want matches when someone is trying to be sly.

So you want something more generic. Using Cabhan's suggestion how about this:

#!/bin/bash -vx

SUBJ="M_ake Money"
tmp="`echo "$SUBJ" | sed "s/[^a-zA-Z ]//g"`"
if [ "$tmp" != "$SUBJ" ]; then echo "Spam"; fi

[thehemi]

I can solve this problem all day w/ perl or bash.
I'm looking for a quick & easy regex answer.

I'm starting to think I might have to abandon the
regex plans and create a custom script instead...

[Cabhan]

You want a regular expression that accepts a normal character followed by any other number of characters, with a non-normal character somewhere in there?

/[a-zA-Z0-9 ]+[^a-zA-Z0-9 ][a-zA-Z0-9 ]*/

This is almost exactly what was in the grep before. grep deals with regular expressions.

[thehemi]

You want a regular expression that accepts a normal character followed by any other number of characters, with a non-normal character somewhere in there?

Not any set of normal characters. Specific phrases that I specify. Make Money, Top Quality, University Degree, whatever the case may be that I'm trying to hunt for the non-normal characters embedded within. I was just curious if there was an easy trick using regex that I was not aware of. I'm starting to think that it's not the case, unfortunately.

[thehemi]

Think of something like this...

M[_\-*]{0,1}a[_\-*]{0,1}k[_\-*]{0,1}e[_\-*]{0,1} [_\-*]{0,1}M[_\-*]{0,1}o[_\-*]{0,1}n[_\-*]{0,1}e[_\-*]{0,1}y

I was just curious if there was, say, a shorthand trick to it.
It's quite obvious that creating and managing many regexes
in this syntax is practically impossible... LOL