In a package in Oracle 19, I'm using:
SDO_LRS.DYNAMIC_SEGMENT (GEOM, START, END, 0.01);
to return a geometry object (SDO_GEOMETRY) from an LRS. The start of the LRS looks like:
-8623852.4702 4852017.8436 0.0,
-8623657.1801 4851950.6596 551.6699,
-8623601.4321 4851930.5192 710.0,
-8623346.1996 4851839.4329 1392.2852,
-8623186.6420 4851783.7162 1817.7965,
...
I've noticed that when clipping and using existing LRS vertices as start and end measures (e.g. 0 and 551.6699), the new geometry matches exactly (green line below is LRS; black is returned segment; left side is the start, right is the end).
When clipping and not using existing LRS vertices (e.g. 42 and 511), it isn't an exact match (green line below is LRS; pink is returned segment; left side is the start, right is the end):
Now we are really just talking about a negligible distance here, within the tolerance, but since I've been asked about it...
Does anyone know how to always get back a perfect match or if that is even possible?
I've tried using SDO_LRS.CLIP_GEOM_SEGMENT instead, but there's no difference (Oracle docs say they are synonyms anyways).
