Algorithm time...

[Glen Bojsza]

Hi Dick,

I tried your solution assuming that plines is my original data in fld
mydata as described.

I get a resulting output incredibly fast but the column 1 values don't
work... ie I think they are putting the line number associated with the
group size and not the original column data?

In other words the two values found in column 2 of the group data must then
get the corresponding column1 values from the original dataset.

For example a basic dataset where column 1 is linear from 1 to 50,000  the
solution should produce only values once from column 1... ie from the first
group of 50 the column 1 values should be two numbers between 1 to 50. The
next group will produce a max and min with their associated values between
51 to 100.

1 13

2 12

3 12

4 13

5 14

6 16

7 14

8 12

9 13

10 12

=====

2    12

6    16

=====

11 10

12 3

13 10

14 11

15 12

16 11

17 13

18 17

19 15

20 16
=====

12    3

17    17
=====
current solution

2    12

6    16

12    3

17    17
All column 1 values will be unique.

Unless I am wrong on what plines is?

Glen