How can I add values per unique string?-CodePudding

Basically, I extract a string and an integer from a text file with:

cat file.txt | cut -d " " -f 1,3 | sort

The output looks like this:

white 3
white 2
white 54
red 23
red 2
yellow 22

How can I add the values per the unique strings with bash scripting?

CodePudding user response：

Sample data file:

$ cat file.txt
white stuff 3 more stuff
white stuff 2 more stuff
white stuff 54 more stuff
red stuff 23 more stuff
red stuff 2 more stuff
yellow stuff 22 more stuff

One awk solution to replace all of the current code and generate totals:

awk '
    { totals[$1] =$3}
END { for (i in totals)
          print i,totals[i]
    }
' file.txt

# or as a one-liner

awk '{totals[$1] =$3} END {for (i in totals) print i,totals[i]}' file.txt

This generates:

red 25
yellow 22
white 59

If sorting is required either pipe the output to sort, eg:

awk '{totals[$1] =$3} END {for (i in totals) print i,totals[i]}' file.txt | sort

Or if using GNU awk:

awk '
    { totals[$1] =$3}
END { PROCINFO["sorted_in"]="@ind_str_asc"
      for (i in totals)
          print i,totals[i]
    }
' file.txt

Both of these generate:

red 25
white 59
yellow 22

CodePudding user response：

First, get out of the UUoC habit.

declare -A totals=()
while read -r color _ count _; do
  totals[$color]=$(( "${totals[$color]:- 0}"   count ))
done < file.txt 
for color in ${!totals[@]}; do echo "$color ${totals[$color]}"; done | sort

output:

red 25
white 59
yellow 22

If it's very big, it would be more efficient in awk.