OK, here are a few test results ... 350,000 points with no additional data
... timings with all data in cache and 17 records selected:
#1
CREATE INDEX "pdata_spatial_index_x" ON "pdata"("gx");
CREATE INDEX "pdata_spatial_index_y" ON "pdata"("gy");
file size 24,200 KB, query time ~6.3 ms
#2
CREATE
On 20 Apr 2016, at 2:16pm, Wolfgang Enzinger wrote:
> CREATE INDEX "pdata_spatial_index" ON "pdata"("gx","gy");
>
> [snip]
>
> However, EQP also shows that it is only used with the value of gx, not gy:
It would be used if you searched for one x value and a range of y values. If
you do have
Simon, Gunter,
thanks for your input. I also noticed that an index on (gx,gy) would help on
covering index searches, however as you both noticed as well there are more
fields in this table. But wait - probably for the first step (which points
are located in the region in question anyway?) it might
An: sqlite-users at mailinglists.sqlite.org
Betreff: [sqlite] storing coordinate pairs efficiently?
Dear group,
since a couple of years now, I'm using SQLite for storing graphic objects like
polygons and lines, with great success. Using the rtree extension makes spatial
queries very efficien
Dear group,
since a couple of years now, I'm using SQLite for storing graphic objects
like polygons and lines, with great success. Using the rtree extension
makes spatial queries very efficient.
In my lastest project, however, these objects are just single points
(coordinate pairs). As I understa
5 matches
Mail list logo