Shell Programming and Scripting

I have a file with a list of 'samples' all in one column and a second file with a list of 'results' for these samples. I am trying to use a FOR loop to go through each sample and gawk the second file to return (the first field of) all the results for that sample.

File 1:
BCM51
CNC11
CNC41
MCW11
UMN51
UWA51
BCM161
BCM211
...
...

File 2:
RS2297516 BCM 5 1 BCM51 A/A C/A FALSE
RS254255 BCM 5 1 BCM51 G/G C/C Mis-Match HOMO
RS254255 CNC 1 1 CNC11 0/0 G/G MISSING
RS1106839 CNC 4 1 CNC41 G/G G/A FALSE
RS2294942 CNC 4 1 CNC41 C/T C/C FALSE
RS3736890 CNC 4 1 CNC41 G/G G/A FALSE
...
...

Desired Result:
BCM51 RS2297516 RS254255
CNC11 RS254255
CNC41 RS1106839 RS2294942 RS3736890
...
...

Code:

#!/bin/ksh

for line in `cat $1`
do
echo $line
gawk     -v sample=$line '
    BEGIN { printf "%s ", sample }
    sample==$5 { printf "%s ", $1}
    END { printf "\n" }' $2
done

All I am getting is errors so far. I have deduced (via commenting out gawk) that it generates an error for each line in the first file and then there is an error on my gawk statement. I can't make sense out of any of them:

Code:

samples.txt: CNC11:  not found
samples.txt[2]: MCW11:  not found
samples.txt[3]: UMN51:  not found
samples.txt[4]: UWA51:  not found
samples.txt[5]: BCM161:  not found
samples.txt[6]: BCM211:  not found
samples.txt[7]: BCM291:  not found
...
...
...
errorsummary.sh[5]: syntax error at line 7 : `'' unmatched

Hello all,
I realized that I was actually executing the wrong script. I had two versions and was editing one but executing the other. No wonder I kept getting the same error no matter what I changed

Heh. And when you find out about fgrep, you might kill yourself

Hmmm. I take it from your response that fgrep would have provided a much easier way to perform the task, but after looking it up I don't see how. (I found this definition: "use fgrep to find all the lines of a file that contain a particular word.")
Seems that I would have just been doing the same 'for' loop except to fgrep instead of gawk. In fact, there would have been an extra step because that would have just given me the lines, but I'd still need to parse out the first field from the rest of the line.
Am I missing something?

The -f option to grep provides a facility for finding all lines in a file that match any of the strings in the -f file. Historically this was only available with fgrep, but with GNU grep, you can always do it. So

Code:

fgrep -f file1 file2

Granted, it doesn't ensure that its the 5th field. But you can use -w option to make sure that it IS a field.

Code:

nawk '{
if(NR==FNR)
	arr[$1]++
else
	brr[$5]=sprintf("%s %s",brr[$5],$1)
}
END{
for(i in arr)
	if(brr[i]!="")
		print i" "brr[i]
}
' file1 file2

Thanks otheus & summer cherry.
I got it all figured out and learned something while I was at it.