[postgis-users] Optimizing PostGIS/Geoserver schema for huge dataset

[Andy Colson]

On 04/03/2017 04:39 PM, Andrew Gaydos wrote:
> Hi,
>
> Yes, in this case it only returns 9 rows, because my bounding box is
> small.

Ah, that's the thing that cuts down rows the fastest, that should be the thing you query on first, then from the subset pull out streamflow and timeid.  Dunno if its possible to structure the data that way though.

  
> I am getting 4x4=16 tiles - the requests are done in parallel so it's
> not a straight 3x16 seconds calculation.

Ah, parallel, yes, that would answer that question.

>
> Ah yes, duplicate indexes! Thanks for catching that! Would that incur
> a performance hit on database reads?

No, it would slow down insert/delete though.

-Andy