Take the 2-minute tour ×
Geographic Information Systems Stack Exchange is a question and answer site for cartographers, geographers and GIS professionals. It's 100% free, no registration required.

Using "sample" to join raster values onto point data, some of my points get assigned a Null raster value.

The raster does not contain Null values and definitely looks like it has a value at the location of each of these points.

Has anyone encountered this problem before?

Coincidentally, all of the points that fail to be matched to a raster cell are in a horizontal line - could this have something to do with raster tiling and if so, how would I check?

share|improve this question
1  
What does the raster look like when you zoom into that horizontal line to such a degree you can make out the individual cells? –  whuber Jun 17 '13 at 14:50
1  
It looks like the individual cells contain data (they each have a value rather than "no value" or background) –  Sideshow Bob Jun 17 '13 at 14:56
2  
Hmmm... Posting an image of the raster and points overlaid (indicating which points received null values) could help. However, I suspect there may be a miscommunication between you and the software concerning the coordinate systems used by the raster and the points. Another possibility is some hidden limitation in the number of points Sample can handle: how many points are you processing? –  whuber Jun 17 '13 at 15:05
    
@whuber - thanks for these pointers. I would like to post up a picture as you ask but on discovering a workaround I deleted the old data. I have just found that Extract values to points does what I want. Alas I have to run it many times as Extract multiple values to points seems to crash. Hopefully this info will be of use for anyone else encountering the problem. For reference there were quite a lot of points, but not a silly number - around 300,000. –  Sideshow Bob Jun 17 '13 at 15:16
2  
Earlier versions of Spatial Analyst had some internal limitations of just under 2^15 = about 32,000 points per operation. I believe it has been cleaned up over the years, but it's worth being a little suspicious of its capabilities. One way to test is to see whether the problems occur with all points at the physical end of the dataset (as it's stored on disk). –  whuber Jun 17 '13 at 15:21

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.