Shell Script - Check whether a single character input is uppercase or lowercase or special character

Question

This is my code i have written. I need a simple code using if elif to check whether the character read is an uppercase,lowercase or a special symbol.

echo "enter a char"
read c

if [[ $c == [A-Z] ]];
then
    echo "upper"
elif [[ $c == [a-z] ]];
then
    echo "lower"
else 
    echo "Digit or special symbols!"
fi

The following is the output I received after inputting characters

enter a char
A
upper

enter a char
a
Digit or special symbols!

aravind@bionic-beaver:~/Desktop$ ./1.sh
enter a char
1
Digit or special symbols!

Solution 1

Try using regular expression tests:

read -p "Type a character" c
if [[ "$c" =~ [a-z] ]]; then
    echo "lowercase"
elif [[ "$c" =~ [A-Z] ]]; then
    echo "uppercase"
else
    echo "Non-alphabetic"
fi

Solution 2

Unless you empty $IFS and add the -r option, read reads a line of input in a very special way.

For instance, if the user enters " \x ", with the default value of $IFS, $c will contain x, not what the user entered.

Also [a-z] doesn't match lower case letters, it matches whatever sorts between a and z in the locale (with some variation in behaviour between shell. For instance, with bash, in many locales, that includes the English letters between A and Y). It could even match sequences of characters in some locales and some tools.

Here, you'd probably want:

printf 'Please enter a character: '
IFS= read -r c
case $c in
  ([[:lower:]]) echo lowercase letter;;
  ([[:upper:]]) echo uppercase letter;;
  ([[:alpha:]]) echo neither lower nor uppercase letter;;
  ([[:digit:]]) echo decimal digit;;
  (?) echo any other single character;;
  ("") echo nothing;;
  (*) echo anything else;;
esac

(that syntax being POSIX sh syntax, you don't even need to install bash).

If you wanted to limit it to the English letters (letters from the latin script without diacritics), you'd need to either name them individually:

([abcdefghijklmnopqrstuvwxyz]) echo English lowercase letter;;

Or fix the locale to C with export LC_ALL=C after the read and before the case statement, but then the (?) test would be invalid as it could incorrectly interpret some character as sequence of characters. For instance a UTF-8 é would be seen as two characters in the C locale.

Comments

[Andy Dalton]

Your script works for me as written (i.e., it correctly identifies A as uppercase, a as lowercase, and 1 as a digit or special symbol).

[FortuneCookie]

Wow..But how?? Are you sure you copied the same lines of code without any modification?

[Andy Dalton]

I copy/pasted your code without modification into /tmp/ex, then ran bash /tmp/ex

[ilkkachu]

Regardless of the fact that the locale setting will change the behaviour here, I do wonder how it's possible that a wouldn't be matched by [a-z], since it's listed as one of the endpoints. Unless you have extra spaces or other characters there, of course.

[FortuneCookie]

Anyway to implement the same using checking the ascii ranges like 65-90 for upper case and 97-122 for lowercase? or can you help with the same code itself as it doesnt validate the lowercase characters.

[FortuneCookie]

can you help me with the same code itself as it doesnt validate only the lowercase characters.? (without using regular expression)

[ilkkachu]

Well, there's also shopt -s globasciiranges in Bash, but of course it still doesn't make [a-z] match an ä.