Can brace expansions and grep be used together in linux - linux

Question

I am trying to count how many times each ASCII printable character is present in a file. I thought a good way to do this might be to list the printable characters in a { } enclosed list, and use grep on each item within the braces. An example code is below. I would like to expand the char list to include all 64 ASCII printable characters. I cannot figure out how to get the code to read and use each characters between the braces separately. I would really like to output a file in the format "character\tcharacterCount". Any suggestions?
char={" ",!,\",#,"\$"}
cat PHRED_scores.txt | grep -e "$char" | wc -m

Answer 1

Below command will display the special characters present in the file and their total count.
grep -oP '[ !\\$#]' file | sort | uniq -c
Explanation:
o - print the match only.
P - grep with Perl-regexp option.
[ !\\&#] - Special characters are included in the character class. You have to escape \ so that it means a literal \
sort Output would be sorted.
uniq -c All the duplicates are counted and then it will be combined into one.

Answer 2

There is a way to avoid listing all 64 characters individually to match the ASCII character set. Bash provides character classes and allows ranges to represent numerous characters without listing each individual character. Some examples are:
[a-z] match all lowercase characters
[A-Z] match all uppercase characters
[0-9] match all digits
[[:print:]] all printable characters
So with very little effort, you can match all upper and lowercase characters and all digits with:
[a-zA-Z0-9]
You can then add the additional printable characters, but you must take care to escape or avoid those with special meaning to regular expressions themselves. An example (not intended to be all-inclusive is)
[a-zA-Z0-0:;~!##$%&*()_-+=]
or you can use the predefined class:
[:print:]
You can add as required. To solve your problem, as Avinash provided sort | uniq -c can provide the individual count. Adding an additional call to wc -m will provide the total. With that, it is not difficult to develop a script that will take the filename as an argument and give the total and individual character counts you require. Something similar to the following will work:
#!/bin/bash
echo -n "Total character count: "
grep $cclass "$1" | wc -m # obtain the total character count
echo -e " Individual frequency:"
grep -o [[:print:]] "$1" | sort | uniq -c # obtain the individual frequency
exit 0
Sample output:
Total character count: 455
Individual frequency:
6 =
10 _
7 -
4 ,
12 ;
1 /
4 .
6 "
9 (
9 )
2 {
2 }
2 *
5 \
2 #
4 %
4 0
3 a
17 b
11 c
1 C
24 d
4 D
28 e
1 E
...