CODEWITHSUNDEEP
bashshellawk
Daniel Luci
Daniel Luci

Reputation: 77

How do I sum the rows of two files with each other? (1 column of data in each) in bash

For example,

I have two files:

file1.log:

123
456
789

and

file2.log:

123
456
789

and I would like to sum them together quickly?

(My actual files have over a million lines each, but equal number of lines.)

The output would be:

246
912
1578

Upvotes: 3

Views: 238

Answers (6)

Chris Koknat
Chris Koknat

Reputation: 3451

Here's a Perl solution:

perl -lne 'BEGIN{open $in1,shift; open $in2,shift} while($n1=<$in1> and $n2=<$in2>){print $n1+$n2}' file1.log file2.log

A more readable (but slightly slower) version:

open my $in1, "<", shift;
open my $in2, "<", shift;
while ( my $n1 = <$in1> and my $n2 = <$in2> ) {
    print $n1 + $n2
    print "\n";
}
close($in1);
close($in2);

Just for kicks, I benchmarked the above vs all other solutions listed here.
They all produce the same output.
The input files are lists of integers from 1 to 999999
seq 999999 > file1.log ; seq 999999 > file2.log

Benchmark results (10 runs each):

shell@thor                     0.070/s
paste@cyrus                    0.410/s
awk@anubhava                   0.546/s
perl_5.6.1                     1.06/s
paste@veerendra                1.32/s
perl_5.20 (readable version)   1.37/s
awk@karakfa (awk 3.1.5)        1.44/s
perl_5.20                      1.75/s

Upvotes: 0

Veerendra Gadekar
Veerendra Gadekar

Reputation: 4472

This could be another option

paste file1.log file2.log | awk '{print $1 + $2;}'

#246
#912
#1578

Upvotes: 1

Thor
Thor

Reputation: 47099

I think paste and bc are the best solution here. Just for fun here is one way to do it with pure bash (adapted from this post on unix.sx):

while read n1 <&3 && read n2 <&4; do
  echo $((n1 + n2))
done 3<file1.log 4<file2.log

Or use the -u fd option with read (thanks rici):

while read -u3 n1 && read -u4 n2; do
  echo $((n1 + n2))
done 3<file1.log 4<file2.log

Output:

246
912
1578

Upvotes: 1

karakfa
karakfa

Reputation: 67467

awk to the rescue

awk '{getline t < "file2.log"; print $0+t}' file1.log

Upvotes: 3

Cyrus
Cyrus

Reputation: 88583

Try this:

paste -d + file1.log file2.log | bc

Output:

246
912
1578

Upvotes: 10

anubhava
anubhava

Reputation: 784988

Using awk you can do:

awk 'FNR==NR{a[FNR]=$1; next} {print $1 + a[FNR]}' file1 file2
246
912
1578

Upvotes: 2

Related Questions