Opened 7 years ago
Closed 5 years ago
#1622 closed defect (duplicate)
shapefile support in rasdaman
Reported by: | bbell | Owned by: | bbell |
---|---|---|---|
Priority: | major | Milestone: | 10.0 |
Component: | clipping | Version: | development |
Keywords: | shapefile, clip, polygon, line | Cc: | Dimitar Misev, Peter Baumann, Bang Pham Huu, Vlad Merticariu |
Complexity: | Hard |
Description (last modified by )
We should implement support for processing shapefiles, e.g. a query option such as
'select clip(..., decode($1)) from ...' -f shapefile
At bare minimum, this would require implementing a decode for shapefiles relying on a library for parsing them (shapelib seems the most popular).
Change History (10)
comment:1 by , 7 years ago
Description: | modified (diff) |
---|
comment:2 by , 7 years ago
comment:3 by , 7 years ago
I understand what you're saying, but the shapefiles have WKT encoded multipolygons and multilines, which we could handle without too much difficulty; however, the point being that there is no rasql grid data does suggest shifting the parsing side over to WCS.
Would we not need to have the capabilities in rasql to handle multipolygons and multilines before introducing this into WCS/WCPS?
comment:4 by , 7 years ago
Cc: | added |
---|---|
Component: | undecided → petascope |
comment:5 by , 7 years ago
We would need support in rasql for reading shapefiles (or wkt files) from file. These files can get very large (100s MB), and specifying the polygons directly in the query is impractical.
comment:6 by , 7 years ago
I think normally, shapefile is geo-referenced and it is not only 1 file (https://en.wikipedia.org/wiki/Shapefile), it is up to users to provide the input in WKT (not shapefile) with WCS for example and Petascope translates these data to polygon vertices.
comment:7 by , 7 years ago
I think normally, shapefile is geo-referenced and it is not only 1 file (https://en.wikipedia.org/wiki/Shapefile)
I read the wiki page and as far as I can tell only the .shp file is relevant for rasdaman. The other files are some indexes or attributes/metadata that are of no use for us.
comment:8 by , 7 years ago
Suggest to gain experience in the use of the new clipping functionality that you have implemented, then we can think about extensions like file format support - it goes substantially beyond what our standards do right now. I would like to see the practical use first.
comment:9 by , 7 years ago
Component: | petascope → clipping |
---|
comment:10 by , 5 years ago
Resolution: | → duplicate |
---|---|
Status: | new → closed |
These are different semantic levels: Shapefiles operate on 2D horizontal georeferenced vectors, rasql on nD abstract arrays. No knowledge is available on rasql level on how to translate.
Hence, this would be more appropriate on WCS/WCPS level. Vector/raster clipping actually is a good candidate for a forthcoming OGC WCS extension standard.