I have a file with two fields of numbers that I want to use as input for another program.

The above code does not work, as I think it would take the whole first and second fields as the input for one particular instance of the program 'inputbashangle'. What I want is to get the first two numbers from the first line of the file 'outfailtest', execute 'inputbashangle' with them, then move on to the first two numbers of the second line for all the lines of 'outfailtest'.

How would I do this?

In scripting, commands are executed in serial fashion; in general, a command runs only after the command preceding it terminates (barring some complicated workarounds).

The usual way to proceed is to first build lists of entries you want to work on, holding them in variables or temp files, then use a separate loop to iterate through them.

In order to give you more detailed advice though, I think we'd need to see an example of the contents of the files in question, as well as what you want the actual output to look like. Your current description is a little too vague for me to understand fully.

1 members found this post helpful.

Well ultimately using awk here seems to be a complete waste as it is doing nothing only awkish.
In saying that it could be used like so:
This now tells awk that the items being referenced are awk fields.

As I said thought, it would just be better off either within your existing script or in a bash script to call yours:

1 members found this post helpful.

To follow up, perhaps something like this would work. It's untested though.

1 members found this post helpful.

Note: I was wrong about what was wrong in the OP. Read the end for details of what's actually wrong.

Ok. I have a simulator called '3bodyangle', which takes in two coordinates and outputs many lines of ten or so numbers each. Right now I have two scripts: the first one merely reads two numbers as input for the simulator:
The second one sends the prior script to a remote cluster. This is the script mentioned in my OP.
I have two scripts because the default cluster output has some extraneous information in it (such as what the submitted job's name is, the CPU time, etc) that I do not want. The first script, 'inputbashangle', gives me the raw data. The file 'outfailtest' looks like this (apologies, the formatting is odd when copied and pasted):
This is what I need help with:
Right now, if I execute 'massinputangle' the correct number of jobs is submitted, which is one job per line in the text file 'outfailtest', which 'massinputangle' uses as input. However, I only get one output file with the information from all of the jobs, which is called 'outangle-'. So it seems that the output from each file is not getting sent to a new output file called 'outangle"$1"-"$2", but rather the "$1" and "$2" are getting ignored in the output file's name, so everything gets sent to one file. Executing the first script, 'inputbashangle', by itself gives the correct output naming (ie if A=2 and B=2, then the output file's name is outangle2-2). So it seems that it's the second script, 'massinputangle', that messes something up.

Read the section about pipes in the info gawk file, and use gawk rather that awk. In gawk. your programs would simply be gawk '{print $1 " " $2 | "./inputbashangle"}' outfailtest

(This use of a pipe is, if I correctly recall the documentation, a gawk extension. Bi-directional pipes are also available if needed.)

Note: I think your code would, perhaps, have worked if you remembered that $1 $2 concatenates those two strings into a single string. Try awk '{c="./inputbashangle " $1 " " $2;system(c)}' outfailtest

1 members found this post helpful.

@PTrenholme - knew there was something I forgot in my haste ... spaces

generate your statements first

1 members found this post helpful.

This worked. Thank you very much.

Thank you to everyone who replied; I shall try to figure out why your suggestions worked while mine did not.

yours did not work because the forming up of your string is wrong.
you should put spaces in between.