Shell Programming and Scripting

this is probably a simple thing but i am new to shell scripting and am having a difficulty.

I need to check a users input to verify that it only canains alphabetic characters and spaces, i.e, no numbers, hyphens, slashes, any other special characters, etc.

at the moment I have:
if [ "`echo \"$input\" | grep [:alpha:]`" ] ... (goes on to do what i want it to do)

this however does not work as any input with at least 1 alphabetic character still is accepted.

any help with this is greatly accepted and if you want to call me a noob i completely understand.

thanks in advance

The following will provide the ASCII representations for every unique character in a text string.

Code:

> echo "abcA1Ahello9%1" | od -An -t dC -w1 | grep -v "*" | sort -u
   10
   37
   49
   57
   65
   97
   98
   99
  101
  104
  108
  111

From the listing, then check for the character codes that you do not want.

What you want to do is check that there are no characters which are not alpha.

Code:

if ! [ "`echo "$input" | grep -v '[[:alpha:]]' >/dev/null`" ]; then ...

Note also the correct way to write [[:alpha:]] with double brackets around it. The redirect to /dev/null is because grep will of course print any matching lines, which in this case you probably don't want. (You could also use grep -q if your grep offers that.)

The backticks inside test inside if are a bit of an antipattern; it comes a lot more elegantly with the use of case, like this:

Code:

case $input in *[!A-Za-z]*) die horribly;; esac

If your shell offers proper POSIX character classes then by all means use [![:alpha:]] instead of the locale-hardcoded [!A-Za-z]