way to do subtraction between two text files in linux

Question

I've two big files, a.txt and b.txt

each contains a lot of entries like this:

say a.txt has these:

12 entry1  
23 entry2  
34 entry4  
...

b.txt has these:

14 entry1  
25 entry3  
14 entry4  
...

how can I do the subtraction for the same entry, while leaving the uniq entries untouched? e.g.

a.txt - b.txt would give me

entry1: 12-14 = -2  
entry2: 23  
entry3:    25  
entry4: 34-14 = 20  
...  

I couldn't figure it out, please help. I know awk should be the way to go, but just cannot figure it out.

Thanks

Answer 1

This is a version without awk, it could help you while you finish your awk script:

#!/bin/bash
#substraction.sh
file1="a.txt"
file2="b.txt"
echo "">myTempFile.txt
fileTemp="myTempFile.txt"

lin1=$(wc -l $file1 | cut -d " " -f 1)
lin2=$(wc -l $file2 | cut -d " " -f 1)

if [ $lin1 -lt $lin2 ]; then
    lineas=$lin2
    myfileBIG=$file2
    myfileTINY=$file1
else
    lineas=$lin1
    myfileBIG=$file1
    myfileTINY=$file2
fi

ind=1

echo "Total of lines: "$lineas
#### myfileBIG with myfileTINY
while [ $ind -le $lineas ] #less than or equal to
    do
        mynum1=$(sed -n "$ind"p $myfileBIG | cut -d " " -f 1)
        mydat1=$(sed -n "$ind"p $myfileBIG | cut -d " " -f 2)
        mynum2=$(cat $myfileTINY | grep $mydat1 | cut -d " " -f 1)
#echo "$mynum1**$mydat1**$mynum2"
        if [ -n "$mynum2" ]; then
            result=$(($mynum1-$mynum2))
            echo "$mydat1: $mynum1-$mynum2=$result"
            echo "$mydat1: $mynum1-$mynum2=$result">>myTempFile.txt
        else
            result=""
            echo "$mydat1: $mynum1"
            echo "$mydat1: $mynum1">>myTempFile.txt
        fi

        ind=$(($ind+1))
        #echo "Posicion: $ind de $lineas"
    done

#### myfileTINY with myTempFile
echo "Using temp file..."
ind=1
if [ $lineas -eq $lin1 ]; then
    lineas=$lin2
else
    lineas=$lin1
fi
echo "Total of lines: "$lineas

while [ $ind -le $lineas ]
    do
        mynum1=$(sed -n "$ind"p $myfileTINY | cut -d " " -f 1)
        mydat1=$(sed -n "$ind"p $myfileTINY | cut -d " " -f 2)
        mynum2=$(cat $fileTemp | grep $mydat1 | cut -d " " -f 1)
#echo "$mynum1**$mydat1**$mynum2"
        if [ -z "$mynum2" ]; then
            echo "$mydat1:    $mynum1"
            echo "$mydat1:    $mynum1">>myTempFile.txt
        fi

        ind=$(($ind+1))
        #echo "Posicion: $ind de $lineas"
    done
echo "...Done!"
echo "Result:"
sort -b myTempFile.txt

Execute follow this:

bash ./substraction.sh

Result: