Read file from within AWK and save $1 to a variable

Shell Programming and Scripting
1

Hi

I am very new to NAWK programming so this question is probably going to sound really stupid:

I have a NAWK script which contains a DO loop. During each loop it runs a FORTRAN program which in turn generates two output files , each one containing 2 integer variables. I would appreciate it very much if someone could tell me how I can read in these two files and save the two integer numbers to variables so I can use them later on in the NAWK script? Thanks in advance.

Cheers

Robbie

2 
getline myVar1 < path2FortranOutputFile1
getline myVar2 < path2FortranOutputFile2

or better yet... - post the relative piece(s) of your AWK script to fully understand the context of the question.

3

Thanks very much for the quick reply.

I was wondering if anyone knows how I could you the getline function to read in each record and store in separate variables?

Cheers

4

Forgot to add the code:

......
system("rm runfile")
print "choose_lib",dbfile,libfile,"outf1","outf2" icoord,jcoord > "runfile"
close("runfile")
system("chmod +rwx runfile")
system("runfile")
system("rm runfile")
......

Basically I am trying to run the "choose_lib" program which is a FORTRAN program that takes 6 arguments. It then dumps two integer values to a file called "outf1" and a further 2 two integers to another file called "outf2".

5

post the relative AWK code and the 'mnemonic' code of what you want to do.
also the output of your fortran runs would help.

6

Sorry, I seem to be writing the message after you reply!

The output from the FORTRAN file is as followed:

" 41 21 "

(excluding the quote marks)

7

Firstly,
you don't need the intermediate 'runfile' - you can do it all natively in 'awk'

Secondly,
you need to close the FieldDescriptors of the commands you run - you might run out of the FDs allocated to awk - this number of the FDs differs from awk to awk, but is limited [6 for Solaris' 'awk'.

Here's something to try [I use single-quoted parameters to the executable just in case if any of them have embedded spaces]:
nawk -v q="'" -f myAWKscript.awk myInputFile

myAWKscript.awk:

BEGIN {
   fortranEXE="choose_lib"
   outf1="outf1"
   outf2="outf2"
}
................
................
cmd=fortranEXE " " dbfile " " q libfile q " " q outf1 q " " q outf2 q " " q icoord q " " q icoord q
cmd ; close(cmd)
getline var11 var12 < outf1
getline var21 var22 < outf2
8

ooops sorry - should have tested it first!

BEGIN {
   fortranEXE="choose_lib"
   outf1="outf1"
   outf2="outf2"
}
{
   ................
   ................
   cmd=fortranEXE " " dbfile " " q libfile q " " q outf1 q " " q outf2 q " " q icoord q " " q icoord q
   cmd ; close(cmd)
   getline var1 < outf1
   getline var2 < outf2

   n1=split(var1, var1A, FS)
   n1=split(var2, var2A, FS)

   var11=var1A[1]
   var12=var1A[2]
   var21=var2A[1]
   var22=var2A[2]
   printf("var1->[%s] var2->[%s]\n", var1, var2)
   printf("var11->[%s] var12->[%s]\n", var11, var12)
   printf("var21->[%s] var22->[%s]\n", var21, var22)
   ................
   ................
}
9

Wow! I'm impressed.

Thanks alot - that certainly did the trick. I appreciate very much your expertise.