[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
|
Re: [udig-users] Oracle: sdo_aggr_mbr vs user_sdo_geom_metadata
|
On 19/08/2010, at 10:34 PM, Bergenroth, Brandon wrote:
> I'm not so worried about uDig not using user_sdo_geom_metadata as much as it doing the full table scans to determine the mbr. And then, doing the query on all tables even before I've picked one to actually display.
As indicated I think I can fix that; we will just have to go with a different assumption (consider bounds based on crs extent if available).
> I'm a little surprised about the negative feedback about user_sdo_geom_metadata, it seems to me the users with problems having correct metadata would be the same users with problems having a correct crs!
I was not trying to be negative - just reporting what I remember of the discussion. Since that time I have found https://jira.codehaus.org/browse/GEOT-704 so you can read for yourself.
> In fact, in my experience when loading data with a 3rd party tool like ESRI ArcCatalog, the metadata is inserted correctly with the proper bounds, but it's the crs (srid) that always shows up null.
Oh cool! Do you have access to a log? Perhaps we could sort out what ESRI catalog is using to determine the bounds?
> And Simon's solution suffers a bit as well, an index is not required, I might only want to display a particular layer at full extent and an index is just a waste in that case, but that is probably a rare occurrence compared to everything else.
>
> I guess I've just always seen user_sdo_geom_metadata as a necessary part of having tables with sdo_geometry, so I make sure it is always populated correctly (same with geometry_columns in postgis).
That is my understanding as well; as you go through the oracle docs they make it very clear that proper set up is required for the index to be efficient.
Thanks Brandon your conversation has already turned up a few good ideas,
Jody
