awk + if ?

Serg · #1 (**permalink**) 12-13-04, 10:00

Hi all.

I have been working to find a solution for this problem but
it is beyond my actual capabilities at this point. If you could
at least say whether this is possible or not I will be very thankful.
I wonder whether awk may be used or not...someone suggested
perl but I prefer to stick with awk.

Here is how the raw data is defined. There are N columns (let's say
8) like this (I add the labels A-H just to make things clear)

A B C D E F G H
-13.1057 3.2476 -9.4736 1 41 251 6 3.7889
-8.2136 -1.2834 -7.6682 1 44 251 8 3.9791
-6.5500 -3.0177 -8.1940 1 99 251 9 4.5311
-1.0685 0.0483 -1.40046 1 144 251 3 3.3905
-17.3394 3.3012 -5.5183 2 18 251 4 3.2886
-13.7365 3.9043 -7.7776 2 41 251 6 3.3796
-18.7815 0.6845 -4.5467 3 18 251 4 3.9160
-13.0219 3.4592 -7.7978 3 41 251 1 3.5433
-8.0730 -4.3466 -6.3149 3 99 251 9 3.8411
-12.4067 -1.4076 -10.1356 3 144 251 3 3.2097
-12.9859 3.0015 -8.7701 4 41 251 1 3.3707
-8.0811 -2.0228 -4.0008 4 44 251 7 4.8075
-8.5788 -3.8426 -7.8051 4 99 251 9 4.0662
-11.9980 -1.7409 -9.9496 4 144 251 3 3.2608
... . .. .. .. ...

The relevant columns are the fourth (D) and the last (H).
What I am looking for is someway to print only the
lines that contain the lowest value of H for each set of
numbers D. In this case,

D H
-1.0685 0.0483 -1.40046 1 144 251 3 3.3905
-17.3394 3.3012 -5.5183 2 18 251 4 3.2886
-12.4067 -1.4076 -10.1356 3 144 251 3 3.2097
-11.9980 -1.7409 -9.9496 4 144 251 3 3.2608

This would be much easier if the total number of lines
for each set D is known but this is not the case.

There are many other problems that I solved using your help
and previous posts around. At this time I know that I need to use
two "if"s. One to filter column D and another to find the lowest
value of H. I am stuck in the second conditional.

Thanks for any suggestion,

Serg

vgersh99 · #2 (**permalink**) 12-13-04, 10:15

nawk -f serg.awk myInput.txt

here's serg.awk:

Code:

BEGIN {
   FLD_d=4
   FLD_h=8
}

!($FLD_d in arr) { arr[$FLD_d] = $0; next }

{
  split(arr[$FLD_d], tmp, FS);
  if ( $FLD_h <= tmp[FLD_h] )
    arr[$FLD_d] = $0;
}

END {
   for (i in arr)
      print arr[i];
}

Serg · #3 (**permalink**) 12-13-04, 10:54

Fantastic, Vlad !
It works very well. I will need to run a 'sort' command to rearrange
the lines because the fourth column D must be in crescent
order. This is no big deal, considering what your script is capable of.

Thanks a lot !

Serg.

--------------------------------------------------------------
-11.9980 -1.7409 -9.9496 4 144 251 3 3.2608
-1.0685 0.0483 -1.40046 1 144 251 3 3.3905
-17.3394 3.3012 -5.5183 2 18 251 4 3.2886
-12.4067 -1.4076 -10.1356 3 144 251 3 3.2097

sort -n +3 output > answer

-1.0685 0.0483 -1.40046 1 144 251 3 3.3905
-17.3394 3.3012 -5.5183 2 18 251 4 3.2886
-12.4067 -1.4076 -10.1356 3 144 251 3 3.2097
-11.9980 -1.7409 -9.9496 4 144 251 3 3.2608

Internet Services

Web Development

Digital Marketing

Consumer Tech