Shell Programming and Scripting

Let me try one more time: I understand that your files' names change daily. I understand that you can't hard code the name of the files to be compared because you want to select a different pair of files to be compared each day. I assume that there could be lots of files in the /tmp/test directory.

Please explain how you select the two files you want to compare out of all of the files that are present in /tmp/test.

I have a question.
e.g.
Assuming that there are 10 lines in the file_A and 5 lines in the file_B, and the content shown below:
file_A file_B
please tell me what's your desired output?

Don: i have only 2 files in that folder...these 2 files need to compare and generate output to other folder.

Frank: with that example, i have to get only "B05" as output.

Code:

diff -c file1 file2 | gawk '/^+/ {print $2}' >> file_diff

However, I feel it's wired. Maybe I have no idea about your real need.

If Aggr_New_06_15_2013_1944.txt and Aggr_New_06_19_2013_1944.txt are the only two files in this directory with names ending in .txt, running the following grep command:

Code:

grep -vF -f *.txt

produces the output you want. You can redirect the output to go wherever you want it. Of course if the two files had names that sorted in the opposite order, the results would be completely different; and, you still haven't specified how to determine which file comes first in the comparison.

Hope this helps,
Don

Thanks Don, its working...but its taking long time compare to "comm" command.
till now we are using manually comm -2 -3 file1 file2 > file3
This command working faster than grep...our file size is large.

I'm very glad that comm is working for you. And, you're very lucky that it is. The behavior of comm is only defined if both input files are sorted in collating order in the current locale. Neither of your sample input files are sorted, so if comm is producing the output you wanted, it is coincidence.

If you had said that your input files were large and sorted, I would have considered comm before fgrep; but since your sample input files were small and unsorted, I used grep -F instead of sorting both input files and then using comm on the resulting sorted files.

You still haven't specified how you decide which of the files in your directory is supposed to be "file1" and which is supposed to be "file2" in your example above. So, it is also just luck that *.txt happened to give the results you wanted when the results would have been very different if the order of the operands had been reversed.

So, with your real input files, is the command sequence:

Code:

for i in *.txt
do      sort -o $i $i
done
comm -2 -3 *.txt > SomeFile_In_SomeOtherDirectory

still faster than:

Code:

grep -vFf *.txt > SomeFile_In_SomeOtherDirectory

Both of the above command sequences should produce the same output.