I have to 'morf' estimated line-fragments between endpoints in a network. I know where the endpoints have to go, so making control-vertices and feeding those to the rubbersheeter exactly does the trick if I feed it one feature.
Feeding it a whole lot of features and controlvertices makes a mess, as a lot of network-repairs are spatially grouped. With many FME transformers the group by (a unique identifier) solves such problems, but the rubbersheeter (and the AffineWarper) don't have this.
Calculating the values for a scaler, ofsetter and rotator (= affiner) is quite a hassle giving the things I 'know' at the desired point in the workflow, while the rubbersheeter does this in a natural, smooth way.
Does anyone have a neat solution for getting the effect of a group-by option in a RubberSheeter?
Kind regards,
Martin
