Shell Programming and Scripting

How to find total number of special character in a column?
I am using

Code:

 awk -f "," '$col_number "[^0-9]*$" {print $col_number}' file.csv|wc -l

but its not giving correct output. It's giving output as 1 even though i give no special character?

Moderator's Comments:

Please use code tags next time for your code and data. Thanks

What is a "special character"?

What is $col_number?

Hi,
$col_number can be anything say 11 or 13. I need the total count of all the special character that can appear in a string. Say, for example in Column 12 i have entries 123.345%65_34,0.345-987,567$9098. I know i can use file separator (awk -F "|" $12(col_number) "[^0-9]*$" {print $col_number}' file.csv|wc -l) any other special character appeneded to file separator. My requirement is to get total number of special character appearing in $col_number say column 12. How can i do it using awk command.
Also, how to get total count alpha numeric in a column say 0.3456wef54,983.rty,trew34.456.

Thanks,
Ajit

So, $col_number is a shell variable? That won't work in awk, awk is not shell. None of your awk code appears like it would work at all.

And a "special" character is any character that's not a number? Okay.

wc -l does not give you the count of characters, it gives you the count of lines. We don't need it here in any case. Strip all non 0-9 characters out of the column with gsub, then print the length of the column. To get alphanumeric, strip out all non-alphanumeric and get the length of that. etc.

Code:

awk -F, '$(COL) ~ /[^0-9]/ { gsub(/[0-9]/, "", $(COL));  print length($(COL)) }' COL=12 filename

or perhaps

Code:

awk -F, '$(COL) ~ /[^0-9]/ { gsub(/[0-9]/, "", $(COL));  print length($(COL)) }' COL="$col_number" filename

The more painstakingness you apply phrasing your request, the better the answers/solutions you get. It always helps to supply sample date, both in and out.
Me, I am not sure I understood what you really need. You did not, for example, tell us what you consider "special characters", although asked to do so.

Try

Code:

awk '{print gsub(/[0-9a-zA-Z]/, "&"), length}' file
25 32

This is for $0, the entire line. Of its 32 chars, 25 are alphanumeric. Apply it to one single field if you want to narrow it down.