RE: multiple sorting

Subject: RE: multiple sorting
From: David_Marston@xxxxxxxxx
Date: Thu, 17 Aug 2000 14:45:27 -0400
I had considered all the spec citations that Jeni provided before I
sent my prior message warning about the borderline cases, so I
consider Jeni's follow-up to be an elaboration of my points, up to
the final paragraph. At the end, she poses some questions.

>Or is the ambiguity that you were referring to about how to deal
>with sorting that includes empty strings and NaNs and whether
>to accept these as sort keys?

It's more about a scenario where the primary sort is numeric, the
secondary sort is alphabetic, and the primary sort key has a mix
of genuine numbers and NaNs. The secondary sort should break the
mass tie among the NaNs, but you can't (yet) count on that bunch
being before all the numbered ones. (The bunch may be last, or it
may vary with the "order" setting.)

I can see the benefits of allowing the NaN items to at least
participate in the sort, as you've already exemplified in your
sorting technique. If allowed, then they should of course compare
equal, and hence all the NaN items should come through in
document order. I think I'll go do some testing of this
scenario.
.................David Marston


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list


Current Thread