[teiid-issues] [JBoss JIRA] (TEIID-2384) Managing Spatial Data Types

[ https://issues.jboss.org/browse/TEIID-2384?page=com.atlassian.jira.plugin... ] Tom Arnold edited comment on TEIID-2384 at 1/22/15 12:53 AM: ------------------------------------------------------------- Oracle and PostGIS will reproject (ST_Transform) in the case where the SRIDs do not match (from docs). For other sources I think we should error, with an eye towards implementing ST_Transform as a system function: # Add ST_Transform as a CAN_PUSHDOWN with an implementation that just says "reprojection not currently supported". Add as supported function for Oracle & PostGIS. # In mismatched-SRID situations, wrap geometry literal in ST_Transform to do conversion to target SRID. (Dependency on storing SRID in column metadata for this to work.) # Change ST_Transform system function to actually do reprojection. I will spend some time in GeoTools trying to get a better understanding of how this works. was (Author: tom9729): (I don't completely understand what you're saying here.) I went with an int (as opposed to Integer) for the SRID because that's how it is on the JTS Geometry. GeometryType has a constant (UNKNOWN_SRID) that indicates that the SRID is not specified. For all of the WKB functions I've looked at, the SRID is an optional argument so I changed JDBC translator to not specify it if it's unknown. I'm not sure if this is necessary if it works as-is, but I think it makes the intent a little clearer for anyone debugging the SQL. https://github.com/Tom9729/teiid/commit/98e3ade67e3561fb91b2a5ab4a2807458... I think this would be a good thing to add in several places. Oracle and presumably Postgres will reproject (ST_Transform) in the case where both sides have an SRID but the SRIDs do not match. MySQL doesn't care because it doesn't use SRID. Thinking ahead, MongoDB only supports WGS84. It would be cool to be able to offer that reprojection capability in non-pushdown cases. I will spend some time in GeoTools trying to get a better understanding of how this works. -- This message was sent by Atlassian JIRA (v6.3.11#6341)