[SOLVED] How to create a batch script for running awk file with trace in NS2?

rimonece · 11-11-2015, 05:58 AM

I have a tcl file including the queue-limit of a particular link. I want to vary this queue-limit(20 times) and run by the same awk file.

Code:

$ns simplex-link $eNB $UE($i) 10Mb 2ms LTEQueue/DLAirQueue 
$ns queue-limit $eNB $UE($i) 100

In Shell, I am running

Code:

 /ns2-35/ ns 22.tcl > 22.txt; gawk -f 22.awk 22.tr

How to create a batch script for this one and run? Is it possible to get the output straight export to an excel file?

joe_2000 · 11-13-2015, 03:32 PM

Whatever command works in the shell, you can just put it into a textfile that starts with

Code:

#!/bin/sh

make that executable with

Code:

chmod +x /path/to/file

and run it. Your command uses relative paths, so either you'll want to change it to use absolute paths or put your script into the same directory as the awk scripts and run it there.

If you want the command to run twenty times you could either
a) put the command 20 times into your script (very unelegant, but maybe a good place to start for a newbie)
b) look into the syntax of loops in shell scripts.

As far as outputting to excel format: No simple way that I am aware of, but depending on the output of your command you could probably output as csv, which can then be imported to excel. Maybe your command is already doing that (impossible for me to tell as I don't know ns and can't see your awk scripts), so if you want to pipe the result of your command to a text file you can use a so-called redirection operator (> or >>):

Code:

command > outputfile.csv

Depending on the redirection operator you will

> : create a new output file, overwriting an existing one if any

>> : append to an existing output file if it already exists or create it otherwise

One last side note: The first line of the script I mentioned above gives you a basic shell, which you can be reasonably sure will be available on any linux distro out there. For more complex scripts it's sometimes useful to use bash instead:

Code:

#!/bin/bash

I use that all the time and find it more convenient to use than plain sh, but it is less portable so some people will recommend against it.

rimonece · 11-13-2015, 06:28 PM

@joe_2000

Thank you so much. I solved my problem by using your direction.

joe_2000 · 11-14-2015, 11:41 AM

Quote:

Originally Posted by rimonece

@joe_2000

Thank you so much. I solved my problem by using your direction.

You are welcome :-)

rimonece · 11-17-2015, 08:20 PM

Procedure:

1. Let, a directory : ( cd/folder_1/code_1) and keep all tcl and awk file there.
2. Open a file in gedit and write the following code:

Code:

#!/bin/bash
# My first script

tcl_file=('10_0.tcl' '10_1.tcl' '10_2.tcl' '10_3.tcl' '10_4.tcl')
tr_file=('10_0.tr' '10_1.tr' '10_2.tr' '10_3.tr' '10_4.tr')
awk_file=('10_0.awk' '10_1.awk' '10_2.awk' '10_3.awk' '10_4.awk')
txt_file=('10_0.txt' '10_1.txt' '10_2.txt' '10_3.txt' '10_4.txt')


for i in {0..4}
do
	echo 'Start==================Start'$i;
	ns ${tcl_file[$i]} > ${txt_file[$i]}	
	gawk -f ${awk_file[$i]} ${tr_file[$i]} 
	echo 'End====================End'$i;
done

3. Save as mmr.sh in a directory ( cd/folder_1/code_1)
4. run : chmod 777 mmr.sh
5. From that directory, run: ./mmr.sh