How to add a Column of Numbers in Bash?

Bash provides several methods to add up numeric columns in data files. This article explores different approaches including awk, loops, and text processing commands, comparing their performance on column summation tasks.

Using the awk Tool

The awk command is the most straightforward approach for column arithmetic. It reads each line, accumulates values, and displays the final sum.

$ awk '{Total=Total+$1} END{print "Total is: " Total}' numbers.csv

Total is: 49471228

To measure performance, use the time command:

$ time awk '{Total=Total+$1} END{print "Total is: " Total}' numbers.csv

Total is: 49471228

real    0m0.228s
user    0m0.141s
sys     0m0.047s

Multiple Columns with Field Separator

For CSV files with multiple columns, use the -F option to specify the delimiter and reference the desired column number:

$ cat prices.csv

Books,40
Bag,70
Dress,80
Box,10

$ awk -F "," '{Total=Total+$2} END{print "Total is: " Total}' prices.csv

Total is: 200

Handling Header Lines

When files contain header rows, use NR!=1 to skip the first line:

$ cat prices.csv

Item,Value
Books,40
Bag,70
Dress,80
Box,10

$ awk -F "," 'NR!=1{Total=Total+$2} END{print "Total is: " Total}' prices.csv

Total is: 200

Using Bash Loops

Bash loops provide an alternative approach, though typically with different performance characteristics.

With expr Command

The expr command performs arithmetic but creates significant overhead due to subprocess calls:

$ time (sum=0; for number in `cat numbers.csv`; do sum=`expr $sum + $number`; done; echo "Total is: $sum")

Total is: 49471228

real    212m48.418s
user    7m19.375s
sys     145m48.203s

With Arithmetic Expansion

Arithmetic expansion $((...)) provides much better performance than expr:

$ time (sum=0; for number in `cat numbers.csv`; do sum=$((sum+number)); done; echo "Total is: $sum")

Total is: 49471228

real    0m1.961s
user    0m1.813s
sys     0m0.125s

Using bc Command with Text Processing

The bc calculator can handle expressions by converting column data into addition statements.

With paste Command

The paste command joins lines with a specified delimiter:

$ cat numbers.csv | head -10 | paste -sd+ -

2+44+6+15+23+0+15+88+82+1

$ time echo "Total is: $(cat numbers.csv | paste -sd+ - | bc)"

Total is: 49471228

real    0m0.244s
user    0m0.203s
sys     0m0.063s

With tr Command

The tr (translate) command replaces newlines with plus signs, requiring a trailing zero for proper bc input:

$ time ((cat numbers.csv | tr "<br>" "+" ; echo "0") | bc)

49471228

real    0m0.217s
user    0m0.203s
sys     0m0.031s

With sed Command

The sed command provides similar functionality using substitution patterns:

$ time ((cat numbers.csv | sed -z 's#<br>#+#g' ; echo "0") | bc)

49471228

real    0m0.343s
user    0m0.281s
sys     0m0.109s

Performance Comparison

Conclusion

For adding columns of numbers in bash, awk provides the best balance of simplicity and performance. The tr + bc combination offers slightly better speed, while bash loops with expr should be avoided due to extreme overhead. Choose awk for readability or text processing commands with bc for maximum performance.