[SOLVED] Character counting

TKH · 08-01-2012, 11:45 AM

Hi, Fellow LQers

I'm trying to make a simple program which can count the number of certain characters (i.e. backslash, dash, or underscore).
However, I don't know what program should I use, let alone what to type. Can someone help me?

Thank you

Wim Sturkenboom · 08-01-2012, 12:06 PM

Some possible ideas here : http://www.unix.com/shell-programmin...cters-sed.html

David the H. · 08-01-2012, 12:28 PM

To count the number of dashes:

Code:

$ text='foo-bar-baz'
$ printf '%s' "$text" | tr -dc '-' | wc -c
2

$ text='foo-bar-baz'
$ char=${text//[^-]}
$ echo ${#char}
2

The first uses standard coreutils, and should be portable across systems.

The second uses a parameter substitution that's only available to advanced shells like bash or ksh. On the plus side, it's all built into the shell, and therefore very fast.

The concept is the same in both cases. Remove everything except the characters you're interested in, and count whatever remains.

danielbmartin · 08-01-2012, 02:38 PM

This may do the trick...

Code:

sed 's/./&\n/g' $InFile |sort|uniq -c > $OutFile

Daniel B. Martin

TKH · 08-03-2012, 12:50 AM

Thank you for your helps, it worked!

grail · 08-03-2012, 02:05 AM

How many ways to skin this puppy:

Code:

text='foo-bar-baz'

grep -o '-' <<<$text | wc -l

awk -F- '{print NF-1}' <<<$text

ruby -ne 'puts $_.scan(/-/).length' <<<$text