Script to compare files recursively using sdiff

[optimusprime]

Hi All,

I have been surfing to get some idea on how to compare same files from two different paths.

one path will have oldfiles directory and another path will have newfiles directory. Each main directories will have sub-directories in them and

each sub-directories inturn will have *.txt files(simple plain text file having several lines in them).


Note : one advantage is that :

All the sub-directories name in both oldfiles and newfiles main directory will have same directory names.

All the text filenames(*.txt) in sub-directories in both main folders will be the same as well

Now,

a. script has to accept 2 paths ie, oldfiles and newfiles directory path as arguments

b. read first sub-folder name from oldfiles and search the same sub-folder name in newfiles path. If same sub folder found, then

c. check, if text files present, if present check if both filename matches, if so, then

d. Sort the files first, then do sdiff on those 2 files and store the results as a seperate file.

So, to give example how folder structure will look like :

Main folders:

oldfiles path : /tmp/oldfiles/

newfiles path : /tmp/newfiles/


Each main folders will have sub-folders :

oldfiles :
subdirA
subA.txt

subdirB
subB.txt

subdirC
subC.txt

newfiles :

subdirA
subA.txt

subdirB
subB.txt



Each sub-dirs will ahve *.txt having same filename in them.


From the above ex :

script should generate sdiff results in output folders as:

subdirA_subA_result.txt
subdirB_subB_result.txt

I hope have mentioned the what i tend to achieve clearly.

from the below script which i wrote it doesn;t checks for same sub-folders/files and even not generating seperate result files instead

it reads all the *.txt files and just produce one single result file.

#!/bin/bash 
 
# cmp_dir - program to compare two directories 
 
# Check for required arguments 
if [ $# -ne 2 ]; then 
    echo "usage: $0 directory_1 directory_2" 1>&2 
    exit 1 
fi 
 
# Make sure both arguments are directories 
if [ ! -d $1 ]; then 
    echo "$1 is not a directory!" 1>&2 
    exit 1 
fi 
 
if [ ! -d $2 ]; then 
    echo "$2 is not a directory!" 1>&2 
    exit 1 
fi 
 
# Process each file in directory_1, comparing it to directory_2 
find $1/ -name '*.txt' -print | while read src 
do 
#for filename in $1/*.txt; do 
#echo $filename 
    fn=$(basename "$filename") 
    if [ -f "$filename" ]; then 
        #if [ ! -f "$2/$fn" ]; then 
            #echo "$fn is missing from $2" 
            #missing=$((missing + 1)) 
        #fi 
                sort $filename 
                #echo $filename 
                sort $2/$fn 
                #echo $2/$fn 
                sdiff $filename $2/$fn | egrep '>|<|\|' > resultfile.txt 
    fi 
#done 
done 
echo "File comparision done, please see resultfile"
 

[Luuk]

It will only create one result because you did not add the path in line 37 like this:

sdiff $filename $2/$fn | egrep '>|<|\|' > $1/resultfile.txt 

NOTE: On a second run of this script, your script WILL find the result.txt of the previous run

NOTE2: The check in line 28

if [ -f "$filename" ]; then 

can be skipped if you change line 23 to:

find $1/ -type f -name '*.txt' -print | while read src