Shell Programming and Scripting

@Corona688,

Yes it does work and yes it does capture the data.

Thank you.

So why not try the solution you built around it, after fixing the capitalized words?

If it doesn't work, tell us exactly in what way it doesn't work, and we'll try to fix it.

By solution you mean the AWK code, also what do you mean by "after fixing the capitalised words?" do you mean the headings for the outputs? Or what?

Still I am not sure steps that I need to do in order to get this working.

I mean the words I told you not to capitalize: Does the input refer to the fifo? And have you tried?

We cannot tell you why it won't work if you won't even try. We don't have your daemon handy. We can't test this.

I think I am confusing myself and confusing you with me too.

This what I have so far

Files

Code:

  data.ksh --> which contains data

Code:

  29203 --> file which I use to capture the data from data.ksh

Code:

  Prw-r- -r - - outputfile

I could read the data from 29203 file through outputfile.

If I cat 29203 > outputfile then from second terminal cat outputfile then I could refer to the data.

Other than that I can't refer to the data using fifo file directly and reading about fifo it says that size of pipe is always zero as it does not store data.

One more thing , now if I refer the data through fifo which is basically reading data from the 29203 file then wouldn't this mean fifo will process the file from the beginning every time the script runs. If yes then it will mean I will have duplicate results.

You have code that looks like it could work if you tried it but refuse to try it. So you have a .ksh file which is not a ksh script?

Created how? So this 29203 is the fifo? Yes you can. 'cat' is not special, cat just reads files. awk also just reads files. Why not try it, like I've asked for days?

Just awk ... /path/to/pipe ought to read one batch of data from the pipe and write it to standard output. Redirect it wherever you please with >filename (to replace) or >>filename (to append). Put it in a 'while true' loop like you already did. each time you run awk it will get exactly one batch of data.

In other words run the code I already got and suggested fixes to, the way you had it originally, except reading from the fifo. And actually try it to see what it does.

Please try it, or at least something. I can't see your computer from here and ran out of things to suggest days ago, we need actual feedback from you, not 'but I think x'.

This file was created using this code (your code)

Code:

while true
do
        cat /path/to/output
done > logfile

I am sure if I am following what you saying , I have tried everything you said.

Code:

 awk ... /path/to/pipe

this will not only get one batch of the data but each time it will read the same data again and again.

I can't read from fifi as I said before using

Code:

mkfifo /path/to/outputfile

Then using awk read from data file to fifo file.
Created a pipe and the size of pipe is always zero as it does not store data. I have tried to write to this file but can't write to the file.

I have tried everything you said and try to explain the results I got.

---------- Post updated 4th Jan 2012 at 05:14 PM ---------- Previous update was 3rd Jan 2012 at 11:40 PM ----------

Please ignore my last reply.

Going back to the pipe, now I have a pipe output file as shows below

Code:

prw-r--r--*** 0 Jan* 3 11:00 OutPutFile

Now I am also able to read from the data file and capture the data using the pipe.

Also using AWK code I can read from pipe to another file which stores the data. (Parsing xml)

But this continues process for example if there are 40 records in the pipe then the AWK code loops again and again without stopping and adding to the counts every time.

I tried using sleep 20 at the end oTf the code before done, but this doesn't make any difference.

Could anyone help with this?

Also if the connection between the pipe and the data file fails, will pipe keep unprocessed data be kept by the pipe until it gets process or will it clear the file?

All I want the script to do is to run for 2 minutes, do calculation for those 2 minutes then sleep for 20 seconds. After 20 seconds start again from the where it was last.