Changes between Version 33 and Version 34 of WKTRasterTutorial01
- Timestamp:
- 06/11/10 10:40:55 (15 years ago)
Legend:
- Unmodified
- Added
- Removed
- Modified
-
WKTRasterTutorial01
v33 v34 128 128 }}} 129 129 130 As you can see, unlike most raster coverage loaded into a GIS, the area covered by this single table raster coverage is not strictly rectangular and the two big missing square areas are not filled with nodata values. Here you have only one table for the whole coverage which is build from many raster files loaded in a single step. You don't need 13 tables to store 13 rasters. We did this with one band SRTM files in TIF format, but it could have been 3 bands BMP files of JPEG and the total size of the coverage does not really matter (PostgreSQL has a limit of 32 Terabytes)...130 As you can see, unlike most raster coverage loaded into a GIS, the area covered by this single table raster coverage is not strictly rectangular and the two big missing square areas are not filled with nodata values. Here you have only one table for the whole coverage which is build from many raster files loaded in a single step. You don't need 13 tables to store 13 rasters. We did this with one band SRTM files in TIF format, but it could have been 3 bands BMP files of JPEG and the total size of the coverage does not really matter (PostgreSQL has a limit of 32 terabytes)... 131 131 132 132 If you want to go to the pixel level and verify the integrity of the values associated to a sample of the raster, (thanks to Jorge Arevalo and GDAL) you can view a vectorization of one of the tile (vectorizing all the table would be way too long) in OpenJUMP. Just type the following SQL query in the same query dialog:
