It doesn't matter where you start.

You can, as you drive "downwards" the records, also drive "upwards" the records.

I am thinking of an algorithm that picks an arbitrary record in the...

john.arnott (4/9/2009)

---

Peso,

I'm not sure that I'm following your toll-booth analogy properly. I don't see how it fits with the data you posted as coming from the OP:

...

Well, I have thought out a proper use-case.

Think of the eight columns as a freeway with 8 lanes.

All 8 lanes are always open.

At the "end" of the freeway (starting from...

I agree with most points, but I strongly disagree with creating sample data that fulfills OP's wanted output. I don't believe in some divine intervention that sample data is perfectly...

I think that sample data is adjusted to OP needs.

OP fourth post (first on second page) still doesn't answer the question of "3, NULL, 9" sample data. OP just rephrase...

Touché !

🙂

But isn't the CorrectSequence based on your understanding of OP's intention?

All assumptions this far seem to be based on page 1 decimal sample data.

That sample data is already sorted! OP writes "The question is how to sort unordered rows into...

I don't get it?

Why is that invalid sample data? Columns are ascending from left to right.

What OP needs is an algorithm to sort the records.

All assumptions this far is based...

Chris, with the logic of the other Chris, how should this sample data be sorted?

1 7 9

NULL 6 ...

Ramesh (4/8/2009)

---

See these results, the row 5 should be followed after row 4 as the col2 value of row 5 is higher than that of row 4.

-- Actual Results

ID col1...

Ramesh (4/8/2009)

---

See these results, the row 5 should be followed after row 4 as the col2 value of row 5 is higher than that of row 4.

-- Actual Results

ID col1...

For Number < 40000 I get these results

Ramesh' Version : 2683 ms

Peso : 1546 ms

Same as before.

And I am little suspicous about your algorithms. For both you and Chris all 6250 records in the final resultset are ordered by original ID.

My suggestion are unordered...

Using Ramesh's test cases, I get these timings:

Ramesh' Version : 526 ms

ChrisM' Version 3 ("condensed" version 2) : 23480 ms

Peso : 236 ms