Aaron Toponce

So, working on one of the many Bash scripts at work, I needed one that handled file manipulation. Not only did it need to manipulate a file, I needed it to handle the manipulation in a loop, as the lines are all the same except for a number (like a primary key) on each line.

Initially, I tried binding vim and emacs keyboard commands to edit the file in the Bash script, then I came to the realization that it would be easier to delete the already existing file, and create a new one from scratch. Here is an example of the file that I would like to create:

marketloc=51.3
maxmarkets=99
market 001,1,7
market 002,2,7
market 003,3,7
market 004,4,7
market 005,5,7
market 006,6,7
market 007,7,7
market 008,8,7
market 009,9,7
market 010,10,7
go
p9999
sample1.lod
quit
quit

The "markets" could go anywhere from 1 to 99 in this case, or whatever "maxmarkets" is set to (above 100, if needed). The first couple lines of the files I need to put in the file before the loop, and the last five lines after the loop. Everything else in the file should be handled in the loop.

First, the scripts needs to be dynamic. From project to project, the number of markets could change, so I need to get this number from the user. Also, the project number, denoted by "p9999", will need to be from the user as well. Both of these arguments I can pass through the command line. We'll call the script "newsamp.sh". Here would be the code, with testing the number of command line arguments (both are required):

Fairly straight forward. Nothing fancy here. We're just making sure that the first argument is the project number, and the second argument is the number of markets. We can do a test to make sure that the first first argument is a valid project number, and the second is a valid range for a market number, but that goes beyond the point of this post.

Now that we have our assignments, lets create the file, and generate the first couple of lines in the file:

Notice I am using 'echo' to add the text to the file. I could use 'cat' or 'sed' or a number of other text options. Here, 'echo' will fulfill our purposes. Notice also that I am doing some I/O redirection using '>>'. I want to append the text to the last line of the file, rather than overwrite the file itself with every 'echo' instance. Now the loop:

There is a couple of spots where we need to do some checking. The first is making sure that the counter is not more than the variable "mkt_numb". If it is, we don't need to do any looping and appending to the file. The second check is to make sure that after the word "market" in our text file, our number is precisely 3 digits, no more and no less. Again, we could do further checking that the number of markets the user enters at the command line is not larger than 999, but that goes beyond the point of this post. So, if the counter is between 1 and 9, then we want to place two zeros before the market number. If it is between 10 and 99, the one zero will suffice, otherwise, print the number. Of course, at every pass, we are using 'echo' to append the text to the file, and we are incrementing our counter by one. The only thing left is to append the last 5 lines to the text file. We want to do this after we have finished with the loop.

Here would be the completed bash script:

That's it! Really pretty easy when you think about it, and accomplished with only a little amount of code. Make newsamp.sh executable, and run the script.

Any questions? 🙂