mjhjmtu

The files contain: File1

1 a
2 b
3 c
...

-------------------File2

2 a
2 c
2 d
...

awk receives this format, seperated by comma:

1 a,2 a
2 b,2 c
...

paste -d, File1 File2 | awk -F, 'n=NF/2;s=""; for (i=1;i<=n;i++)printf "%s%s", s, ($i-$(i+n));s=",";; print ""'

I use this line to do a difference row by row of the first field;
First row being 1-2 so prints -1

Ok it's all good I get the right answer, still I don't see how it can pick up the number field when the field is seperated by -F, representing

1 a

If I change

1 a 

to

a 1

doing the same with the rest it doesn't work which I agree in logic.
So, in other words, how does it get the number when the field is the number and letter? $i = number, why not letter also?

how does it get the number when the field is the number and letter?

When , is your field separator (-F,), text like 2 a is a single field, and it is simply treated as the number 2 when it is used as a number.

The AWK language is very permissive when you ask it to treat a string as a number. It reads the string until the point where it stops making sense as a number. At that point, if there was something that made sense, that's your number. Otherwise it takes it as 0. This all happens automatically when you use the string in an arithmetic expression. To see how it works, run:

awk 'print $0 + 0'

$0 is the whole line. Adding zero forces it to be interpreted as a number but doesn't change what number that is. Then you can type in lines of input to see how awk treats them as numbers. You will find, for example, that the empty string is 0, 3a is 3, 1 2 is 1, 3.4.5.6 is 3.4, and so forth. (Press Ctrl+D at the beginning of a line to signal end-of-input and quit awk, or just press Ctrl+C.)

Depending on what you actually want to do, you may be able to write a simpler command that is more clear about what it does and how it works, and that can be modified more easily. For example, if each file has just two columns separated by whitespace, and the entries do not themselves contain whitespace, just merge them into a four-column table by running paste without -d (or, if you prefer, with something like -d ' ') and process the table by piping it to awk without -F:

paste File1 File2 | awk 'print $1 - $3'

Then it becomes obvious how you would adapt the AWK script when the numbers appear after the letters instead of before:

paste File3 File4 | awk 'print $2 - $4'

(To test this, you can make File3 and File4 from File1 and File2, respectively, by using awk to switch the columns. For example, awk 'print $2, $1' File1 > File3 makes File3.)