UNIX for Dummies Questions & Answers

diff -C999999 tends to give you all the data of both files. How is that not suitable for processing?

@DGPickett
In post #1 the expected output from the process was one line - the line containing "3".


On my version of "diff" we get quite a lot of unwanted output from "diff -C 999999". Admittedly it is possible to fish out the differences by writing a program to unravel the output from "diff".

Code:

diff -C 999999 file1 file2
*** file1    Tue Feb  1 17:09:30 2011
--- file2    Tue Feb  1 17:10:16 2011
***************
*** 1,3 ****
--- 1,4 ----
  1
  1
  2
+ 3

There are much much better methods. The only way that I can explain that my post #4 did not work is if these are not unix text files. Another explanation is that this is a MAC or some weird Linux like Busybox.

No fishing, the diffs start with +/-, but of course if the records are not bound by a symbol and run across many lines, the keys needed to process the diffs are hard to deal with. Now, if the key fields are sometime before and sometimes after the diffs, life gets seriously twisty. It seems like the files need to be parsed into a rectangular normal matrix of rows and columns, and then compared.

Hmm. Just realised that in post #7 the O/P might have run "comm" with no parameters.
I don't think that my post #4 has been run by the O/P.
We really need some representative data because the O/P mentioned that these are actually logfiles and has not answered questions about the order of the data. From the reaction to "diff" the files are not proper unix text files ... or maybe they are active logs?