Using an awk script, how do I count the number of fields in which a substring lies in a file?

Question

For example, I have the text document:

"Hello, I am the janitor and I have a headache
 The rabbit jumped over the red brick wall"

I want to count the number of fields in which the substring 'he' occurs. I do not want to count the number of 'he' in the file, just the number of fields.

So for my example, it should print something like this out:

Number of fields that contain 'he' in record #1: 3
Number of fields that contain 'he' in record #2: 2

It has to be an awk script.

Answer 1

The fields are numbered starting from 1, and NF contains the number of them. So we can iterate over them with for (i = 1; i <= NF; i++) { ... }, with $i inside the loop referring to the field in question. (i is actually just the number of the field, we need the $ operator to get the actual contents of the field.)

And as for finding if a value contains a particular substring, using a regex is easiest. s ~ /foo/ would see if the variable s matches the string foo anywhere, i.e. if it contains it as a substring. Now, you may also want to also match an uppercase letter, in which case e.g. [Ff] would work in place of the f. The bracket group [...] matches any one of the characters inside.

Of course you'll also need a counter, but that's easy, just initialize a variable to zero before the loop (e.g. count=0) and increment if there's a match (count += 1).

~ ~

So, basically, an awk script to run some code for each line/record of a file is just

awk '{ some code }' < filename.txt

Inside the code block, the for loop fits, and it also takes a block in braces { .. }.

awk '{ for ( ... ) { some code } }`

And an if works similarly,

if (condition) { some code... }

(They actually look just like for and if in C.)

And you can use semicolons to separate statements, so

awk '{ what to do before the loop; for ( ... ) { some code }; what to do after }`

Answer 2

Done by below awk script

awk -v i="he" '{print "Number of fields that contain" " " i " " gsub("he",$0) " " "in record " NR}' file

output

Number of fields that contain he 3 in record 1
Number of fields that contain he 2 in record 2