LinuxQuestions.org - awk puzzle

- Linux - General (https://www.linuxquestions.org/questions/linux-general-1/)

- - awk puzzle (https://www.linuxquestions.org/questions/linux-general-1/awk-puzzle-831022/)

Hi, I am struggling to write an awk script with the correct syntax and wondered if anyone knows this really well.

I have 2 columns of data like this below which I would like to addition.

229K 128M
2.41G 4G
126M 2G
2.86G 0
2.69G 4G
175M 2G
2.86G none

1. Convert none to 0
2. If mb(M) then do nothing. If kb(K), or gb(G) then convert to mb
3. Addition the 1st and 2nd column and print output in gb

Code:

cat file | sed -e 's/none/0/g' | nawk -F" " '{if($1 | $2 ~ /G/) {printf "%.2f %.2f\n", $1*1024, $2} elif {if($1 | $2 ~ /K/) {printf "%.2f %.2f\n", $1/1024/1024, $2}}' else 'print $1, $2' | nawk -F" " '{sum += $1, sum2 =+ $2}END{ printf("Used : %12.fGB Quota : %12.fGB\n",sum/1024, sum2/1024)}'

Code:

{

if ( $1 ~ /none/ ) {  sub(/none/, "0", $1) ; arr1=$1}

if ( $2 ~ /none/ ) { sub(/none/, "0", $2); arr2=$2 }

if ( $1 == 0 ) { arr1=$1}

if ( $2 == 0 ) { arr2=$2 }

if ( $1 ~ /M/ ) { sub(/M/, "", $1) ; arr1=$1 }

if ( $2 ~ /M/ ) { sub(/M/, "", $2) ; arr2=$2 }

if ( $1 == 0 ) { arr1=$1}

if ( $2 == 0 ) { arr2=$2}

if ( $1 ~ /G/ ) { sub(/G/, "", $1) ; arr1=$1*1024 }

if ( $1 ~ /K/ ) { sub(/K/, "", $1) ; arr1=$1/1024 }

if ( $2 ~ /G/ ) { sub(/G/, "", $2) ; arr2=$2*1024 }

if ( $2 ~ /K/ ) { sub(/K/, "", $2) ; arr2=$2/1024 }

arr3=arr1+arr2

printf "%.2f %.2f %.2f\n", arr1  , arr2  , arr3

}

Code:

0.22 128.00 128.22

2467.84 4096.00 6563.84

126.00 2048.00 2174.00

2928.64 0.00 2928.64

2754.56 4096.00 6850.56

175.00 2048.00 2223.00

2928.64 0.00 2928.64

It's not best way but should work.
Besides it should work only for upper-case M and K.
It can be fixed by tolower function.

Code:

sed 's/^/((/;s/$/))\/1024000/;s/ \+/)+(/g;s/none/0/;s/K/*1024/g;s/G/*1024000/g;s/M//g' file |bc

Hi goldenbarb, nice post thanks for the awk, that's something that I can re-use later on.

However I have to addition the total of column 1 and the total of column 2 instead of adding up column1 and column2, and outputting this into column 3.

I tried using the following, but the numbers didnt add up.

sum1+=$arr1
sum2+=$arr2

Hi,

Building on goldenbarb's code from post #2:

Code:

#!/bin/bash



awk '{ if ( $1 ~ /none/ ) {  sub(/none/, "0", $1) ; arr1=$1}

if ( $2 ~ /none/ ) { sub(/none/, "0", $2); arr2=$2 }

if ( $1 == 0 ) { arr1=$1}

if ( $2 == 0 ) { arr2=$2 }

if ( $1 ~ /M/ ) { sub(/M/, "", $1) ; arr1=$1 }

if ( $2 ~ /M/ ) { sub(/M/, "", $2) ; arr2=$2 }

if ( $1 == 0 ) { arr1=$1}

if ( $2 == 0 ) { arr2=$2}

if ( $1 ~ /G/ ) { sub(/G/, "", $1) ; arr1=$1*1024 }

if ( $1 ~ /K/ ) { sub(/K/, "", $1) ; arr1=$1/1024 }

if ( $2 ~ /G/ ) { sub(/G/, "", $2) ; arr2=$2*1024 }

if ( $2 ~ /K/ ) { sub(/K/, "", $2) ; arr2=$2/1024 }

arr3=arr1+arr2

col1=col1+arr1

col2=col2+arr2

col3=col3+arr3

printf "%10.2f%10.2f%10.2f\n", arr1, arr2, arr3 }

END { print "--------------------------------" ;

printf "%9.3fG%9.3fG%9.3fG\n", col1/1024, col2/1024, col3/1024 }' infile

I'm not sure if you want to add both the rows and the columns, the above does both. The sum of the columns are shown in G.
The bold parts are the additions and/or changes.

Hope this helps.

Code:

{

if ( $1 ~ /none/ ) {  sub(/none/, "0", $1) ; arr1=$1}

if ( $2 ~ /none/ ) { sub(/none/, "0", $2); arr2=$2 }



if ( $1 == 0 ) { arr1=$1}

if ( $2 == 0 ) { arr2=$2 }



if ( $1 ~ /M/ ) { sub(/M/, "", $1) ; arr1=$1 }

if ( $2 ~ /M/ ) { sub(/M/, "", $2) ; arr2=$2 }



if ( $1 == 0 ) { arr1=$1}

if ( $2 == 0 ) { arr2=$2}



if ( $1 ~ /G/ ) { sub(/G/, "", $1) ; arr1=$1*1024 }

if ( $1 ~ /K/ ) { sub(/K/, "", $1) ; arr1=$1/1024 }



if ( $2 ~ /G/ ) { sub(/G/, "", $2) ; arr2=$2*1024 }

if ( $2 ~ /K/ ) { sub(/K/, "", $2) ; arr2=$2/1024 }



sum1=sum1+arr1

sum2=sum2+arr2

printf "%.2f %.2f\n", arr1  , arr2

}

END{print "------"; printf "%.2f %.2f\n", sum1  , sum2}

Code:

0.22 128.00

2467.84 4096.00

126.00 2048.00

2928.64 0.00

2754.56 4096.00

175.00 2048.00

2928.64 0.00

------

11380.90 12416.00

Yet another way :)