Re: Index on points

2010/9/25 Tom Lane <tgl@sss.pgh.pa.us>:
> Jeff Davis <pgsql@j-davis.com> writes:
>> There's no reason that there couldn't be a point <@ box operator in the
>> opclass, but nobody really uses these geometric types that come with
>> core postgres (at least, not that I can tell).
>
> Actually, as of 9.0 there is a point_ops opclass for GIST, with these
> indexable operators:
>
>  >^(point,point)
>  <<(point,point)
>  >>(point,point)
>  <^(point,point)
>  ~=(point,point)
>  <@(point,box)
>  <@(point,polygon)
>  <@(point,circle)
>
> I agree that for any more than light-duty geometric work, you ought
> to look at PostGIS.
>
>                        regards, tom lane

Thank you Jeff for your reply, that solved the problem.

Tom, would you like to elaborate on that  PostGIS should be used for
other than "light-duty" geometric work?
Is it speed, accuracy or features that is the difference?
For this project I think <@(point,box) is sufficient. What would it
take to motivate a switch to PostGIS for that?

Best wishes.