Hi list,
forward Antonio's reply. It seems that has problems
with his corporative mail ...
>>> Antonio Redondo Lora 03/10/07 9:41 >>>
Thanks Jesse,
I'm surprised too beacuse it often works in 1GB computer,
but, as I said, it only fails (throwing Java Heap Exceptions) in
several laptops.
Indeed, applications operates with layers that have lots of
attributes but I think it depends more on the number
of points of each polygonal feature instead of this attributes
number. Despite this errors, Obviously using this kind of
collections lowers times of geometry operations like
intersection or visitors.
Nevertheless, I have update two of our plugins to
the community repo:
\community\dap
One of them shows a dialog in which we can build new layers
from filtering expressions composed with attributes and operators
(es.dap.sig.udig.filter).
The other, creates several statistics from points layers. It groups
points using other polygonal layer which will be classified with a default
theme buildig from one of this stataistic results.
(es.dap.sig.udig.analysis).
Both of them, use IndexedFetureResults.
PD: We go on working at these plugins. Sorry for possible fails or
bad traductions ...
Saludos
Antonio Redondo Lora
Subdirección de Integración de Sistemas
E.P. Desarrollo Agrario y Pesquero
c/Bergantín, 39 - 41012 Sevilla (España)
Email:
arlora@xxxxxxTlfno: 955059762
Corporativo: 359762
>>> Jesse Eichar <
jeichar@xxxxxxxxxxxxxxx> 03/09/07 5:44 >>>
IndexedFeatureResults loads all features into memory that is the
reason that it is both fast and uses lots of memory. I'm surprised
at the amount of memory it takes though... 500 features doesn't seem
like that much... Perhaps they have many attributes. A few things
that can help are good use of Query so that you only load the
attribute you care about into memory. Do you really need
IndexedFeatureResults for the operation? If not then don't load the
geometry and that will lower the memory requirement.
Hope this helps:
Jesse
On 9-Mar-07, at 3:49 AM, Antonio Redondo Lora wrote:
> Hi udiggers ....
>
> Anyone working with IndexedFeatureResults?
>
> We are working with polygonal layers containing near 500 features. We
> try to minimize the access time to calculate several visitors, for
> example AverageVisitor, SumVisitor or MaxVisitor. The lowest time
> is got
> applying IndexedFeatureResults to localize the features of this
> layers.
> Even so, we have noticed that Eclipse executing in computer with near
> 1Gb of memory has no problems but executing fails at the time of
> aplication installing in some computers with the same amount of memory
> (Its throws Java Heap Exception!!).
>
> Obviously, the time we can use habitual FeatureResults or
> FeatureCollection but the time is longer, near double. Is there an
> alternative in Geotools to index or improve the acces to this kind of
> Feature Results?
