Shell Programming and Scripting

Hi,

I have a text file that is around 7Gb which is basically a matrix of numbers (FS is a space and RS is \n). I need the most efficient way of plucking out a number from a specified row and column in the file.

For example, for the value at row 15983, col 26332, I'm currently I'm using:

Code:

x=26332
y=15983

awk -v row=$y -v col=$x 'NR==row {print $col;exit}' big_file_of_numbers.txt

Obviously, the larger the value of x and y (especially y) the longer it takes. A few seconds per value is normal. I need to do this thousands of times.

I'm not sure if I can use NR like this; I mean is one allowed to 'set' the value of NR? It seems to work, it's just not very quick. I'm using exit to prevent it reading the whole thing.

Cheers,

Jon

Simple. You obviously have a file with rows and columns in it. Call it rcfile
Assume the format is

Code:

 13   99
128  50

13 == row number; 99 == column number

Code:

awk 'FILENAME=="rcfile" {arr[$1]=$2}
       FILENAME=="bigfile" && FNR in arr {print $arr[FNR]}'  rcfile bigfile   > newfile

[/code]