I have a file of which I need to count all partial matches for an input string in a file.

I'll show you an easy example of what I need:

In a file with this content:

Good-Black-Cat
Bad-Red-Cat
Bad-Gray-Dog
Good-Golden-Dog
Bad-White-Dog
Good-Tabby-Cat
Bad-Siamese-Cat

I need to count how many times does the partial string "Good -*-Cat" (Where * could be anything, it doesn't matter) appears. The expected output count is 2.

Any help will be appreciated.

Given

$ cat file
Good-Black-Cat
Bad-Red-Cat
Bad-Gray-Dog
Good-Golden-Dog
Bad-White-Dog
Good-Tabby-Cat
Bad-Siamese-Cat

then

$ grep -c 'Good-.*-Cat' file
2

Note that this is a count of matching lines - so for example it won't work for multiple occurrences per line, or for occurrences that span lines.

Alternatively, with awk

awk '/Good-.*-Cat/ n++ END print n' file

If you need to match multiple possible occurrences per line, then I'd suggest perl:

perl -lne '$c += () = /Good-.*?-Cat/g wc -l

If you also need to match occurrences that may span a line boundary, then you can do so in perl by unsetting the record separator (note: this means that that the whole file is slurped into memory) and adding the s regex modifier e.g.

perl -0777 -nE '$c += () = /Good-.*?-Cat/gs { say $c' file

awk, multiple occurences, space-separated

$ awk 'for(i=1;i<=NF;i++ ) count+=match($i,/Good-.*-Cat/);ENDprint count' input.txt
4
$ cat input.txt
Good-Black-Cat
Bad-Red-Cat
Bad-Gray-Dog
Good-Golden-Dog Good-Whatever-Cat Good-Something-Cat
Bad-White-Dog
Good-Tabby-Cat
Bad-Siamese-Cat

sed + wc, non-multiple occurences

This uses negative pattern matching //! with d for delete, leaving only lines of interest.

$ sed '/Good-.*-Cat/!d' input.txt
Good-Black-Cat
Good-Golden-Dog Good-Whatever-Cat
Good-Tabby-Cat
$ sed '/Good-.*-Cat/!d' input.txt | wc -l
3

Shell solution, non-multiple occurences

Here's shell way that combines case...esac and file-reading loop:

$ n=0; while IFS= read -r line || [ -n "$line" ]; do case "$line" in "Good-"*"-Cat") n=$((n+1));; esac; done < input.txt; echo "$n"
2

Or with indientation

n=0
while IFS= read -r line || [ -n "$line" ]; do 
 case "$line" in 
 "Good-"*"-Cat") n=$((n+1));; 
 esac
done < input.txt
echo "$n"

Explanation:

• 
  n=0 initializes n counter variable


• 
  while IFS= read -r line || [ -n "$line" ]; do...done < input.txt is standard file-reading loop used in shell scripting, with || [ -n "$line" ] protection to account for possible files that don't end in newline


• 
  case "$line" in "Good-"*"-Cat") n=$((n+1));; esac pattern-matching for the desired string with $((...)) arithmetic expansion to increment the counter variable.

Non-fancy sed/grep version

sed 's/(Good-[^ ]*-Cat)/XXXXn/g' input.txt | grep -c XXXX

While XXXX can be any pattern that does not appear otherwise in your file. This approach replaces all matches with the XXXX pattern and a newline, so to make it easily countable by a basic grep expression.

By the way if you take "Where * could be anything" literally, at least to my understanding, the output of any such program would always be 0 or 1, so I am assuming that it should not contain a space at least.