tips with awk with changing parameters - awk

Question

i got several pieces of code that look like:
for ff in `seq 3 $nlpN`;
do
npc1[$ff]=`awk 'NR=='$ff' {print $1}' p_walls.raw`;
echo ${npc1[$ff]};
npc2[$ff]=`awk 'NR=='$ff' {print $2}' p_walls.raw`;
npc3[$ff]=`awk 'NR=='$ff' {print $3}' p_walls.raw`;
npRs[$ff]=`awk 'NR=='$ff' {print $4}' p_walls.raw`;
echo $ff
done
as You can see i'm invoking awk several times. Is there a faster way to do this, like invoking awk once and do the assignments with the changin parameters?
thanks a lot in advance!
input looks like:
...
3.76023 0.79528 0.307771 8729.82
3.76024 0.814664 0.307849 8650.2
3.76026 0.845679 0.307978 8802.97
3.76025 0.826293 0.307897 8690.43
3.76017 0.65959 0.30722 8936.07
...
im looking for sth like:
TY

Answer 1

That does look pretty inefficient. As written, awk is processing the input file in its entirety four times with every pass of the loop. I'm also pretty sure that cut is completely unnecessary unless you have the FS environment variable set to something strange. The following will replace the multiple awk runs with a single pass over the data file that stops after it finds the line. Then you can use cut to extract the individual fields.
for ff in `seq 3 $nlpN`
do
data=`awk 'NR=='$ff' { print $1, $2, $3, $4; exit }' p_walls.raw`
npc1[$ff]=`echo "$data" | cut -f1 -d ' '`
echo ${npc1[$ff]}
npc2[$ff]=`echo "$data" | cut -f2 -d ' '`
npc3[$ff]=`echo "$data" | cut -f3 -d ' '`
npRs[$ff]=`echo "$data" | cut -f4 -d ' '`
echo $ff
done
Note that I added an exit statement so that awk will exit after processing the line. This prevents it from reading the entire file on every pass. If all that you need to do is extract a single line from a file, then you might want to use sed instead since (IMHO) the script is easier to read and it seems to be a little faster on large files. The following sed expression is equivalent to the awk line:
data=`sed -n -e "$ff p" -e "$ff q" p_walls.raw`
The -n tells sed to only output from the lines that are selected by the script. In this case, the script, supplied as two -e parameters. Each is an address followed by processing command. Multiple commands are separated newlines in sed scripts but they can also be specified by multiple -e parameters with the same address. Putting this all together, the expression 42 p tells sed to select line 42 and run the p command which prints the selected pattern space (the 42nd line). The 42 q command tells the utility to exit after processing the 42nd line. So, our sed expression reads the first $ffth lines from "p_walls.raw", prints the $ffth one and exits.

Answer 2

Run awk a single time and process the output on each iteration separately.
awk "(NR > 3 && NR <= $nlpN)"' { print NR, $1, $2, $3, $4 }' p_walls.raw |
while read ff c1 c2 c3 Rs
do
npc1[$ff]=$c1
echo ${npc1[$ff]};
npc2[$ff]=$c2
npc3[$ff]=$c3
npRs[$ff]=$Rs
echo $ff
done