UNIX for Dummies Questions & Answers

Hi,

I was a typical Windows guy. Like to do things just by clicking my mouse. I got a new job now...where they are big on unix.
I am trying to wet my fingures now with unix. Haven't taken the dive yet.

I am trying to find a solution for this problem.
Please help me with some ideas.

Now the problem:
For ex: I have 6 files.
input_1.txt
input_2.txt
input_3.txt
input_4.txt
input_5.txt
input_6.txt

I want to write a script to concat the files on the parameter I pass to the script. Say If I pass parameter 2 to the script.
It should do this....
cat input_1.txt input_2.txt input_3.txt > input_1.txt
cat input_4.txt input_5.txt input_6.txt > input_2.txt

So, If I pass 3 to the script.
It should do this....
cat input_1.txt input_2.txt > input_1.txt
cat input_3.txt input_4.txt > input_2.txt
cat input_5.txt input_6.txt > input_3.txt

And the initial files I have can be dynamic....one day...I may have 20 or 30 files. Could be 25 (odd) too. The script should be dynamic enough.
I hope...my problem is clear enough for you.

Thanks for all your time.
-Sandeep

How does the script know how many files you have to deal with? Will that be an additional parameter?

Let's say you have n input files and want j output files. Do you want each outputfile to comprise n/j input files? Do you always want the output file to be named the same as the input file? In the same directory?

Hi,
Please find my answers below.

How does the script know how many files you have to deal with? Will that be an additional parameter?
Yes. Sorry forgot to mention that part.
The script should take 2 parameters.
#1 the no.of input jobs
#2 the no.of output jobs

Let's say you have n input files and want j output files. Do you want each outputfile to comprise n/j input files?
Yes that is what I expect. n/j input file.

Do you always want the output file to be named the same as the input file? In the same directory?
Yes.

Once again, thanks a lot for your time

-Sandeep

Very basic script, no error checking etc etc ect.

Script takes 1 parameter, number of output files needed,

Script assumes all files which should be concatenated start with input_

Script assumes all files are in the current directory.


Should be easy to change the script into something more luxerary, it just presents a concept.

Code:

#!/usr/bin/bash
NUMOUT=${1}

NUMFILES=`ls input* | wc -l`

((IN_PER_OUT=${NUMFILES} / ${NUMOUT} + 1))

CURRENT=1
OUT=1
ls input* | while read FILE
do
  if [ ${CURRENT} -eq 1 ]
  then
    cat ${FILE} > output_${OUT}.txt
  else
    cat ${FILE} >> output_${OUT}.txt
  fi

  ((CURRENT=${CURRENT}+1))
  if [ ${CURRENT} -gt ${IN_PER_OUT} ]
  then
    CURRENT=1
    ((OUT=${OUT}+1))
  fi
done

Thank you, I tried your script.
It is working close to what I need.
I tweaked it a bit to meet my exact requirements.

Thanks again
-Sandeep