Rio Rio - 1 month ago 6
Linux Question

AWK / Linux script calculations from a text file

I wrote an AWK script that read files and multiplies rows by columns and sum them up. I want to pass several files (any number of files) as an argument in awk script at a time e.g. A.txt, B.txt, C.txt. I want the AWK script to give me sum of rows and columns. I always want to skip first 5 columns from each text file.

Each text file can have any number of columns. There can be several text files inside a folder.

I want to run as:

awk -f foo.awk A.txt B.txt C.txt


e.g.

If there are 3 different files A.txt, B.txt, C.txt, sum up multiplication of rows and columns from each 3 files.

Output should be:

No of columns in A.txt: count of columns in A.txt with first 5 columns ignored
No of columns in B.txt: count of columns in B.txt with first 5 columns ignored
No of columns in C.txt: count of columns in C.txt with first 5 columns ignored
Sum of A.txt: rows in A.txt*columns in A.txt
Sum of B.txt: rows in B.txt*columns in B.txt
Sum of C.txt: rows in C.txt*columns in C.txt
Total Sum: A+B+C


Below is (sort of pseudo-code) what I have got so far for foo.awk (it is not working with multiple files):

#!/bin/gawk -f

BEGIN { rows=0; columns=0 }
{
FS="\t";
if(/^#COLS/) {
column=NF-5; #skip first 5 columns
columns+=column
}
if (!/^#/){
rows++;
files[FILENAME]++;
}
}
END {
for (fname in files) {
printf ("%'24d rows in %s\n",files[fname],fname);
}
printf("No of columns in A.txt= %'d\n", columnsA);
printf("No of columns in B.txt= %'d\n", columnsB);
printf("No of columns in C.txt= %'d\n", columnsC);
sum=columns*rows; # multiply no of rows by column in each file and add them up
printf( "Sum of A.txt %d\n", sumA);
printf( "Sum of B.txt %d\n", sumB);
printf( "Sum of C.txt %d\n", sumC);
printf( "Total sum is %d\n", sum_of_A+B+C);
}


e.g.

A.txt:
#ignore this line -- pattern does not match
#ignore this line -- pattern does not match
#COLS A B C D E F G H I
row1 1 2 3 4 5 6 7 8 9
row2 1 3 3 4 5 6 7 8 9
row3 1 3 3 4 5 6 7 8 9

B.txt:
#ignore this line -- pattern does not match
#ignore this line -- pattern does not match
#COLS A B C D E F G H
row1 1 2 3 4 5 6 7 8
row2 5 3 3 4 6 6 7 8
row3 8 3 3 4 5 6 7 8

C.txt:
#ignore this line -- pattern does not match
#ignore this line -- pattern does not match
#COLS A B C D E F G H I J
row1 1 2 3 3 5 6 7 8 9 2
row2 7 3 3 4 5 6 7 8 9 7
row3 9 3 3 4 5 6 7 8 9 6
row4 9 3 3 4 5 6 7 8 9 6

output:

No of columns in A.txt: 5
No of columns in B.txt: 4
No of columns in C.txt: 6
Sum of A.txt: 3*5=15
Sum of B.txt: 3*4=12
Sum of C.txt: 4*6=24
Total Sum: 12+9+20 = 51


Thank you.

Answer

with plain awk you can do this

$ awk '!/^#/{cols[FILENAME]=NF-5; 
             rows[FILENAME]++} 
         END{for(f in cols) print "No of columns in " f, cols[f]; 
             for(f in cols) 
               {r=rows[f];
                c=cols[f];
                sum+=r*c; 
                sumstr=sumstr?sumstr"+"r*c:r*c; 
                print "Sum of "f ":",r "x" c "=" r*c} 
             print "Total Sum: ", sumstr, "=", sum}' {A,B,C}.txt

No of columns in C.txt 6
No of columns in B.txt 4
No of columns in A.txt 5
Sum of C.txt: 4x6=24
Sum of B.txt: 3x4=12
Sum of A.txt: 3x5=15
Total Sum:  24+12+15 = 51

There is a mismatch in number of columns, are you skipping 5 or 6. Also note that the order of entries is not preserved, can be fixed with gawk sorted-in, or with little extra coding as below...

$ awk 'FNR==1{order[++k]=FILENAME} 
        !/^#/{cols[FILENAME]=NF-5; rows[FILENAME]++} 
          END{for(i=1;i<=k;i++) print "No of columns in " order[i], cols[order[i]]; 
              for(i=1;i<=k;i++) {f=order[i];r=rows[f];c=cols[f];sum+=r*c; sumstr=sumstr?sumstr"+"r*c:r*c; print "Sum of "f ":",r "x" c "=" r*c} 
              print "Total Sum: ", sumstr, "=", sum}' {A,B,C}.txt

No of columns in A.txt 5
No of columns in B.txt 4
No of columns in C.txt 6
Sum of A.txt: 3x5=15
Sum of B.txt: 3x4=12
Sum of C.txt: 4x6=24
Total Sum:  15+12+24 = 51