Shell Programming and Scripting

Sorry for the probably strangely worded title but I don't really know how else to put it.

Background context: Post processing LAMMPS simulation data.
tl;dr: I'm making two spheres collide, every defined timestep the simulation outputs a bunch of data including total energy of the particles, their coordinates, etc...

Here is my problem:

I'm using awk to post process this data, what I need right now is for it to find the closest point two values approach, give me the closest approach distance (called overlap in the code since it's actually two particles overlapping slightly) and give me the line number that happens on.

So X1 & X2 will first approach each other, reach a closest point then distance themselves from each other.
Finding the closest distance is not a problem.
I do it thus:
The relative code includes the bits with 'overlap' 'max' 'dx' 'z1' 'z2' etc..

The other code is for a similiar problem: Once the energy reaches equilibrium I want that equilibrium and the line where it happens.
However in this case that's not a problem, since it will be the last line treated so FNR is correct.

Code:

BEGIN{
oldte=0
te2=1
max=0
OFS=","			# Changes Output Field Seperator to a comma (default is space) 
				# so it can be appended to a .csv file which excel opens without a prompt
printf("%s,%s,%s,%s,%s,%s,%s,\n","Change","Et1","Et2","ColDt","dt","steps","max")

}
{
	if((FNR+1)%11==0)
	{
		te1=$5
		z1=$1
		
	}
	
	if(FNR%11==0){
		te2=$5
		dt=$6
		z2=$1
		dx=(z2-z1)
		overlap=(0.05-dx)
		if(overlap>max)
		{
			max=overlap							# overlap calculations
		}
		if(oldte==te2)
		{
			steps=FNR/11
			colDt=dt*steps
			printf("%s,%s,%s,%1.8f,%s,%s,%1.8f\n",change,te1,te2,colDt,dt,steps,max)
			exit														
		}
		else
		{
		oldte=$5
		}
		
	}
		
}
END{

}

So the problem here is that I can't use

Code:

ClosestPoint=FNR/11

in the if expression for finding the max value of overlap because when I call that variable to be printed along with the other data, it will be reevaluated and use the FNR at that point in the input file (which will be to far down so to speak).


It's probably a lot more explanation then needed for my question but I thought I'd give as much context as possible.

I guess it just comes down to this:

Can I set a variable X to equal the value of NFR at the time that X is set and not have X be reevaluated when I print it.

I am having trouble understanding your problem..

In general I can say that is you are processing data and can only find out if a certain line contains the right data later one or more lines later on, then you could record the data into a variable. For example if you want to use printf(...) to print a value, instead you can use some_var=sprintf(...) to record an output line into a variable, and print this recorded variable (print some_var) one or more lines down when you are able to conclude that it was indeed the right value..

Hope this helps and/or makes sense...

So f.e.

Code:

some_var=sprintf(NFR)

and then later

Code:

print(some_var)

Would give me the original value and some_var would not be reevaluated as being =NFR ?

Right. Of course in this example, you would record somevar=FNR , I only used sprintf() as an example so you can record including the formatting.

This works if the proper line is the previous line. If it can be several lines back, you can use a "circular buffer" by using an array,

For example:

A[FNR%4]=sprintf(...) would record the last 4 lines, so you can select which one to print...

Ah yes see that's the problem, the 'overlap' NFR will always be several (100+) lines before the oldte==te2 (energy equilibrium) NFR, because of how the simulation works.

So if the some_var still gets evaluated as =NFR when I print it at the end
f.e. print(some_var) then i'll have the wrong value. On the other hand if it only gets evaluated as being = an integer (a previous NFR in this case) than I'm golden.

An array wouldn't help me since I can't know which line I need, that's the whole point of getting the NFR value at that point.

I'm not at all sure that I understand what you're trying to do, but if you change:

Code:

		if(overlap>max)
		{
			max=overlap							# overlap calculations
		}

to:

Code:

		if(overlap>max)
		{
			max=overlap							# overlap calculations
			max_line=FNR
		}

then when you get to the END section of your awk script, the variable max contains your largest overlap you found and the variable max_line contains the line number in your input file where you found that maximum.

That's the thing I did something similar and it didn't give me the correct FNR instead it seemed to reevaluate the variable (max_line in your example) to =NFR

Although I'm not sure whether I put that printf() in the END section, then it the variable might not get reevaluated