r join on the group_id and compute all
>>>>>
>>>>> In fact you could avoid to create explicitly the cells, but it will be
>>>>> more complicated.
>>>>>
>>>>> Cheers,
>>>>> Rémi-C
>>>>>
&g
Wood
>>>>
>>>>> I figure you have spatially indexed the polygons already?
>>>>>
>>>>> Any way of pre-categorising your polygons - binning them in some way
>>>>> that allows a non spatial test in the where clause to replace the
o indicate if a feature has any part <= a
>>>> particular x value or not.
>>>>
>>>> run this against the 2m features once to populate it, and assuming you
>>>> get a 50/50 split of T/F features, your 2m^2 query can instead include a
>>>>
s
>>> different, they cannot touch, so your query should involve a spatial test
>>> only over 1m^2 instead... if you do this on both x & y, the boolean filter
>>> will replace even more spatial calculations & drop them to 500,000^2
>>> tests...
>>>
&g
ace even more spatial calculations & drop them to 500,000^2
>> tests...
>>
>> If you can pre-classify features so that non-spatial tests can reduce the
>> spatial ones (esp for features with lots of vertices) in a where clause,
>> such queries do run much faster, but you
t; ----------
> *From:* Lee Hachadoorian
> *To:* PostGIS Users
> *Sent:* Wednesday, November 20, 2013 8:52 PM
> *Subject:* [postgis-users] Finding Islands
>
> I am trying to find "islands", polygons in a (multi)polygon layer which
> are not connec
adoorian
To: PostGIS Users
Sent: Wednesday, November 20, 2013 8:52 PM
Subject: [postgis-users] Finding Islands
I am trying to find "islands", polygons in a (multi)polygon layer which
are not connected to any other polygons in the same layer. What I came
up with runs in a couple o
I am trying to find "islands", polygons in a (multi)polygon layer which
are not connected to any other polygons in the same layer. What I came
up with runs in a couple of seconds on a layer with ~1000 geometries,
and a couple of minutes on a layer with ~23,000 geometries, but I want
to run it o
