does the size of the dbf file influence the size of the spatial index (qix)?

[Frank Warmerdam]

On 7/22/05, Bart van den Eijnden <BEN at syncera-itsolutions.nl> wrote:
> Hi list,
> 
> normally my spatial indexes are small compared to the size of my shapefiles..
> 
> Now I have the following situation:
> 
> 04-07-2005  16:46          163.888.026 acn.dbf
> 04-07-2005  16:46           39.730.468 acn.shp
> 04-07-2005  16:46            4.966.396 acn.shx
> 22-07-2005  16:21          119.741.008 acn.qix
> 
> So the spatial index is 119 Mb, while the shapefile is only 39 Mb.
> 
> For another shapefile I have the same type of situation:
> 
> 28-02-2005  10:58           22.805.942 hnrtekst_text.dbf
> 28-02-2005  10:58            1.793.808 hnrtekst_text.shp
> 28-02-2005  10:58              512.588 hnrtekst_text.shx
> 22-07-2005  16:24            6.434.784 hnrtekst_text.qix
> 
> Does the dbf size influence the size of the spatial index? I can't make any sense out of that, but maybe somebody can ... ?

Bart,

Generally speaking the .qix file size is related to the number of features
but not to the size of .dbf (more attributes make no difference).   The
depth you specify when building the .qix file will also have an effect.
If you pick a high depth value, there will be alot of lightly populated
"deep" leaves to the tree. 

Generally speaking point datasets are the worst case, and it might
be helpful to manually pick a modest "depth" (like 4) rather than 
accept the default or pick a big number. 

I have sometimes felt that the .qix files were more bulky than is 
strictly necessary. 

Best regards,
-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | Geospatial Programmer for Rent