ST_DFullyWithin

[Paul Ramsey]

Is built fresh off master?

> On Jan 25, 2024, at 3:45 PM, Bruce Rindahl <bruce.rindahl at gmail.com> wrote:
> 
> " I don't think that word means what you think it means"
> 
> Say A is a point at 0.1,1
> B is a line from 0,0 to 0,2 
> 
> Then using the following code:
> --------------------------------------------------------------
> with a as (select st_makepoint(0.1,1) as geom),
> b as (select st_makeline(st_makepoint(0,0),st_makepoint(0,2)) as geom)
> 
> select 'ST_Covers(ST_Buffer(A,Dist),B)',  ST_Covers(ST_Buffer(a.geom, 0.2), b.geom) from a,b
> union all
> select  'ST_DFullyWithin(A,B,Dist)',  ST_DFullyWithin(a.geom,b.geom,0.2) from a,b
> union all
> select 'ST_Covers(ST_Buffer(B,Dist),A)',  ST_Covers(ST_Buffer(b.geom, 0.2), b.geom) from a,b
> union all
> select  'ST_DFullyWithin(B,A,Dist)',  ST_DFullyWithin(b.geom,a.geom,0.2) from a,b
> -------------------------------------------------------------
> 
> 
> Yields:
> QUERY                                                RESULT
> "ST_Covers(ST_Buffer(A,Dist),B)"      false
> "ST_DFullyWithin(A,B,Dist)"                false
> "ST_Covers(ST_Buffer(B,Dist),A)"      true
> "ST_DFullyWithin(B,A,Dist)"                false
> 
> I think the last row should also be true.
> 
> 
> 
> On Thu, Jan 25, 2024 at 1:14 PM Paul Ramsey <pramsey at cleverelephant.ca <mailto:pramsey at cleverelephant.ca>> wrote:
>> 
>> 
>>> On Jan 25, 2024, at 11:57 AM, Martin Davis <mtnclimb at gmail.com <mailto:mtnclimb at gmail.com>> wrote:
>>> 
>>> Another reason for keeping the meaning "ST_DFullyWithin(A, B) = ST_Covers(ST_Buffer(A, Dist), B)":
>>> 
>>> I see the main use of this function in queries to be "find all the B features which are fully within distance D of an A feature".  So the A feature is the *constant* "query item", and it is evaluated against a set of B features.  It seems more intuitive for the constant feature to be the first argument in functions.
>> 
>> Concur, this is also my intuition, which is why it’s coded the way it is…
>> 
>> P
>> 
>>> 
>>> PS note the definition needs to use Covers rather than Contains, due to that ol' quirky definition of Contains! 
>>> 
>>> On Thu, Jan 25, 2024 at 10:57 AM Paul Ramsey <pramsey at cleverelephant.ca <mailto:pramsey at cleverelephant.ca>> wrote:
>>>> 
>>>> 
>>>>> On Jan 25, 2024, at 10:19 AM, Bruce Rindahl <bruce.rindahl at gmail.com <mailto:bruce.rindahl at gmail.com>> wrote:
>>>>> 
>>>>> I agree with Regina.  Also why are we coding this as a function in C when it can be done either of two ways in SQL:
>>>> 
>>>> In general functions in SQL are less fun to upgrade. Will eventually do this via Hausdorf.
>>>> 
>>>> I don’t know that I agree about the parameter ordering, and I do not think the name of the function provides any guidance when I re-write it in object form, 
>>>> 
>>>> A.DFullyWithin(B,R)
>>>> 
>>>> at least not in the same way that
>>>> 
>>>> A.contains(B) 
>>>> 
>>>> makes parameter meaning clear.
>>>> 
>>>> It’s the “D” that wrecks it.
>> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-devel/attachments/20240125/fc64253f/attachment.htm>