DROP TABLE #T;

CREATE TABLE #T(Col1 INT)

INSERT INTO #T

SELECT 1 UNION ALL

SELECT 1 UNION ALL

SELECT 2 UNION ALL

SELECT 3 UNION ALL

SELECT 4 UNION ALL

SELECT 1

--SELECT SUM(DISTINCT(Col1)) AS 'Sum of distinct values',SUM(Col1)...

Here's the entire SQL set that solves the challenge using negation. SW1.doc and SW2.doc are separated versions of the previous attachment. I can now process the entire set in under...

For those who want to explore this a bit further, I'm attaching the SQL that JC provided to create the 'Sweden' table, and a 'Map' table to associate all old...

FWIW, with some interesting results, I did solve the challenge, but ironically, it was with my original SQL on a screaming database server using the divide and conquer method. What...

>>Les - One other thing. Would you mind posting your final row counts (input and output), CPU and elapsed times to solution?

Will do. It's still running. If it doesn't...

After putting the route issue to bed, I was mulling over how well this performed. So, to escalate its application, I tossed this at the Jan. '98 Dr. Dobbs...

Thanks folks. Brilliant! It made mincemeat out of the Route problem. It worked so well though that I'm now tossing it at a much bigger NP problem that we attempted...

Sheesh...quite the brain trust up here. It's going to take me a bit to wade through these. I'll report the results after I come up for air 🙂

~Les

Chris,

Oddly...I haven't seen that one. Any idea where I might track it down?

~Les

So, the permutation count for David's scenario is 50*49*48*47*46 = 50!/45! = 254,251,200

That would leave 120 (5*4*3*2*1) permutations for each distinct combination, so the combination count = 50!/(45!*5!) = 2,118,760...

>>Is there adjacency data missing from the description?

The adjacencies have already been resolved.

Thx,

~Les

>>And so, a link to Hugo Kornelis's bin packing series, which may spark an idea or 2.

Thanks. It does have similarities. I'll have to give it a thorough read.

Lynn,

It's a redistricting (route) requirement, where we're looking to contain each district to roughly the same number of meters (service locations). Urban growth and all that.

>>All the possible combinations of those rows that added to 5?

All the possible DISTINCT combinations.

~Les

G[]...

I'd come to the same conclusion, but thought it might be worth a toss up here. I thumbed through all of Celko's books, but nothing jumped out. Intuitively I think...