ADQL 2.1: Preferred crossmatch syntax

[Mark Taylor]

On Fri, 3 Nov 2017, Francois-Xavier PINEAU wrote:

> Dear Mark and DAL,
> 
> Although I agree with the approach, there is one point that bothers me.
> 
> I know that not everyone agrees on the matter, but I personally prefer to
> declare the cross-match condition in the JOIN (like in your InteropOct2015DAL
> presentation) rather than in the WHERE.
> 
> In fact, in relational algebra, I prefer to see the cross-match as a
> theta-join, not as a selection on the result of a cross-product.

OK, I don't really have an opinion on that, I never really know about
what syntax is best for joins, or whether it makes a difference.
[And I'm afraid I wouldn't know a theta-join if one trod on my foot]

Is it likely to make a difference in performance terms which of these
two ways it's done?  I.e. do we need to specify here what's the
preferred way to use a constraint like DISTANCE()<threshold,
or is this something we can harmlessly leave to the taste of
the query author?

Mark

--
Mark Taylor   Astronomical Programmer   Physics, Bristol University, UK
m.b.taylor at bris.ac.uk +44-117-9288776  http://www.star.bris.ac.uk/~mbt/