[postgis-devel] [PostGIS] #1005: [raster] Problem with ST_SummaryStats(text, text) results

[PostGIS]

#1005: [raster] Problem with ST_SummaryStats(text, text) results
----------------------------+-----------------------------------------------
 Reporter:  pracine         |       Owner:  pracine      
     Type:  defect          |      Status:  new          
 Priority:  medium          |   Milestone:  PostGIS 2.0.0
Component:  postgis raster  |     Version:  trunk        
 Keywords:                  |  
----------------------------+-----------------------------------------------

Comment(by pracine):

 Replying to [comment:2 dustymugs]:
 > As for the standard deviation of a coverage, I do confirm the issue.  To
 do this correctly will require going over the coverage table twice, once
 for the coverage's mean and the second for the mean's standard deviation.
 So, do we want to be accurate at the expense of speed or be close at the
 expense of accuracy?  Maybe we need to add an option flagging whether or
 not to get the accurate standard deviation for a coverage?

 1) We must be clear about that in the doc and we should provide (a link
 to?) the method used to get the approximative value

 2) I agree with the option. Default should be the approximate value.

 I thought the method you found (one pass stddev) provided accurate result.
 I thought we added the sum to be able to get an accurate stddev. How does
 it work for one tile? Is the stddev accurate for one tile?

-- 
Ticket URL: <http://trac.osgeo.org/postgis/ticket/1005#comment:4>
PostGIS <http://trac.osgeo.org/postgis/>
The PostGIS Trac is used for bug, enhancement & task tracking, a user and developer wiki, and a view into the subversion code repository of PostGIS project.