Hello,
I have a map of questionable quality and I'd like to use the geometry validator. Since there is actually clear warning that this transformer is intended for experienced users, is there a guide of some kind I could refer to?
I did see the help section, but I am looking at some simple how to start ... what are the best ways to set up the transformer, if you have no idea what you are doing etc. Few hints really.
I have attached sample of the file.
Thank you
Best answer by fmelizard
Hi @jitkam, I would start off by using "Basic Integrity" or "OGC". These can be good places to start with to see how the data is behaving.
We also have a data validation tutorial series. Although it might not contain information about the GeometryValidator, it can help you if you're looking to identify certain issues in your data (i.e.- self-intersections, duplicate features, etc.). I would recommend inspecting the data like @jkr_da suggested to start with.
+35I think it's best to start looking at the map and it's questionable regions. And then try to see if the GeometryValidator helps solving the problem.
But be aware that not every 'not-validated' geometry is always a problem. And that validated geometry is without problems.
I think that's why it says it's an advanced user transformer. If you already have a map that is showing the area, this could be fine. But if you need the data to be used in other transformers and there are unexpected results this is a problem. But validating your geometry may not be the solution for your unexpected results. Then you have to write your own validation for this dataset.
But you can use it to detect area's that need extra attention, to make your map better.
+20Hi @jitkam, I would start off by using "Basic Integrity" or "OGC". These can be good places to start with to see how the data is behaving.
We also have a data validation tutorial series. Although it might not contain information about the GeometryValidator, it can help you if you're looking to identify certain issues in your data (i.e.- self-intersections, duplicate features, etc.). I would recommend inspecting the data like @jkr_da suggested to start with.
