Re: Using Generic Comparator in Oracle ORDER BY

And ultimately, I did opt to use a custom PL/SQl function F that translates inputs C1 as follows:

{C1, F(C1)} ==>

{
(A-1, A0001),
(A-1.1, A00010001),
(A-1.2, A00010002),
(A-1.2.1,A000100020001),
(B-1,B0001),
(B-2,B0002),

...
}

Not an altogether bad solution - it essentially translates C1 into a string that can be sorted lexically so

SELECT C1
FROM MY_TABLE
ORDER BY F(C1) gives the desired results.

Notice, however, that my function would fail if I every received "A-10000". It is also computationally more expensive than a direct comparison function would have been. And finally, it just seems rather obfuscous, IMO... it doesn't have the "elegant" feel that I would get from a
"ORDER BY C1 COMPARING WITH F" construct... but then, maybe I'm just weird that way. Received on Wed Oct 26 2005 - 16:03:25 CDT