Bash script to compare numbers in a txt file

leopard86 · 10-18-2007, 09:15 AM

Hi everybody, I've only a basic preparation on shell and I urgently need to have a script to do this:
I have a file with some lines written in log.txt:

Breakpoint 1 at main()...
line 10 printf("some C code"...
Hello I return 23
Program exited with code 025

Now the script must read the last two number (in this case 23 and 25) and tell me wether they are different or not...
I really need some hints, and if possible some peaces of code instead of a general explanation...
Good bye...

radoulov · 10-18-2007, 10:02 AM

In case awk is acceptable:

Code:

awk '/Hello I return/{v=$NF;next}
$NF+0!=v{print "They are different ...";exit}' filename

with bash it would be:

Code:

while IFS= read -r;do
        case $REPLY in
                ("Hello I return"*)v="${REPLY##* }";;
                ("Program exited with code"*)[ "${REPLY##* }" -ne "$v" ]\
                &&echo "They are different";break;;
        esac
done<filename

ktgajowniczek · 10-18-2007, 11:37 AM

if you don't like awk you can also do it like this:

FILENAME=log.txt; if [ $(( `tail -n2 $FILENAME | sed "s/.*\ 0*$[0-9]*$/\1/" | paste -s | sed "s/$[0-9]*$[^0-9]$[0-9]*$/\1-\2/"` )) -eq 0 ]; then echo "equal codes"; else echo "not equal codes"; fi

--
KTG

/bin/bash · 10-19-2007, 08:45 AM

Code:

#!/bin/bash
NUMS=( $(grep  -o "[0-9]\+" temp.log) )
if [ "${NUMS[(( ${#NUMS[@]} -2 ))]}" != "${NUMS[(( ${#NUMS[@]} -1 ))]}" ];then
  echo "They are different"
else
  echo "They are the same"
fi

The first line puts all numbers in the NUMS array.
The test checks the last 2 numbers to determine if they are the same.
NOTE: This (( ${#NUMS[@]} -2 )) means Number of elements in array NUMS - 2.
Remember array elements start a 0 so the last array element is actually (Number of elements in array - 1.)

ghostdog74 · 10-19-2007, 07:39 PM

Quote:

Originally Posted by ktgajowniczek

if you don't like awk you can also do it like this:

FILENAME=log.txt; if [ $(( `tail -n2 $FILENAME | sed "s/.*\ 0*$[0-9]*$/\1/" | paste -s | sed "s/$[0-9]*$[^0-9]$[0-9]*$/\1-\2/"` )) -eq 0 ]; then echo "equal codes"; else echo "not equal codes"; fi

--
KTG

may i suggest you organize your code in a more readable fashion in future. thanks