UNIX for Dummies Questions & Answers

Hi.
I am not sure how to solve this problem and if it is possible to do so with scripting.

Let's say I have this data:

A 12345 12360
A 12359 12380
A 12381 12390
A 12400 12450
A 12451 12460
B 23456 23460
B 23470 23480
B 23477 23505

I wan't each line to be compared in this way: If the values of column 2 and 3 of line 1 overlap values of column 2 and 3 of line 2 then keep first line only.
Then keep comparing.
So my result would be:

A 12345 12360
A 12381 12390
A 12400 12450
A 12451 12460
B 23456 23460
B 23470 23480

Thanks for your input!

Should it consider all previous lines, or only just the previous line?

Only the very last one!

Okay, that's doable:

Code:

awk '(NR>1) && (X2>=$2) && ($3>=X1) { next }; { X1=$2; X2=$3 } 1' data

Which amounts to:

Code:

awk '
# If this isn't the first line, and it intersects the range X1-X2, skip it.
(NR>1) && (X2>=$2) && ($3>=X1) { next }
# record the new range and print
{ X1=$2; X2=$3 } 1' data

Thank you so much.
It seems to be working though I don't really follow how the program is working.
Is it comparing e.g.
Line 2 to Line 1. If there is no overlap, it keeps both lines and then compares line 3 to line 2. Then line 3 overlaps with line 2 so it kicks line 3 out and then compares line 4 to line 2?

---------- Post updated at 03:19 PM ---------- Previous update was at 02:42 PM ----------

Oh yes, and another question: How would the code be different if it were to consider all previous lines.
Thx

It would potentially need to store values for each and every line to compare later.

This could be an awful lot of computational work, n^2 intersection checks for n lines.

---------- Post updated at 12:47 PM ---------- Previous update was at 12:29 PM ----------

Code:

$ cat allset.awk
{       for(N=1; N<=MAX; N++) if(($3 >= X1[N]) && (X2[N]>=$2))  next; }
{       MAX++; X1[MAX]=$2; X2[MAX]=$3; } 1
$ awk -f allset.awk < data
A 12345 12360
A 12381 12390
A 12400 12450
A 12451 12460
B 23456 23460
B 23470 23480
$