Count numbers present in each line for unique String using shell script

Question

I am using unix shell script and have an Input File with data as:

3:abc
1:xyz
1:abc
2:def
10:xyz

My expected output is:

4:abc
11:xyz
2:def

i.e. Find unique string on each line after delimiter and add up the numbers before that. How to do this?

Answer 1

Here is a solution using awk. It accumulates the values into an array.

awk -F ":" '{count[$2]+=$1} END {for (key in count) print key, count[key]}' awk_data.txt

And here is a version, using a bash script:

#!/usr/bin/env bash
declare -A count
while read line; do
    key=${line##*:}
    cnt=${line%%:*}
    count[$key]=$(($cnt + ${count[$key]=0}))
done < "$1"
for K in "${!count[@]}"; do echo $K ${count[$K]}; done

And another bash version from the comments, using IFS=:

#!/usr/bin/env bash
declare -A count
while IFS=: read -r cnt key; do
    count[$key]=$(($cnt + ${count[$key]=0}))
done < "$1"
for K in "${!count[@]}"; do echo $K ${count[$K]}; done

Answer 2

You may also want to take a loot at GNU datamash utility. Example (table.txt from PO):

$ sort -t: -k 2 table.txt | datamash -t: -g 2 sum 1 collapse 1
abc:4:1,3
def:2:2
xyz:11:10,1

• -t: field delimiter is :
• -g 2 group by the (previously sorted) second column
• sum 1 sum the correspondent values in column 1
• collapse 1 and also collapse them in a "," separate list