Question

Clipper doing strange things to .dgn cells

  • 27 March 2013
  • 5 replies
  • 1 view
A
Badge +1
I have a .dgn file with many cells that I am clipping.  I am not editing the cells in any way.

 

 

I read it in, run the clipper (on shapes from another dgn) and write it out, attaching a cell library as necessary.

 

 

Where the output is clipped, there are no cells, obviously.  Where it is not clipped, there are only cells where it is within the bounding box of a clipper, and active points everywhere else.

 

 

I am doing nothing with the clipper's bounding boxes at all.  I am using the clipper in a very simple way.

 

 

I have no idea why cells would only be preserved as cells when they are in the bounding box of a clipper, and dropped to point everywhere else.

 

 

When I inspect the output directly in Data inspector, attributes appear identical, except the cells carry the _fme_no_clipper attribute, and the features that are dropped to points do not (all of these features are outside the clipping boundaries.)

 

 

What is going on?  Is this a feature of the clipper I didn't know about?  A crazy bug?

5 replies

david_r
Rank
Userlevel 4
Hi,

 

 

can you provide us with some screenshots of your Clipper connections and settings, as well as an example of the results you get versus what you want?

 

 

(Upload your screenshots to a picture sharing site and then specify the URL into the Image button dialog)

 

 

It is a bit difficult to go by the description alone.

 

 

David
A
Badge +1
Certainly.

 

 

This is a very large process, with hundreds of thousands of .dgn features and 145 clipping shapes.  I will show a zoom in on a small region.

 

 

I have created screenshots in a Dropbox gallery, here.

 

 

https://www.dropbox.com/sh/d8bzis1q45f681o/MJgI7S7iBm#/

 

 

Clipper Parameters - This is the clipper transformer attributes window.

 

 

Microstation Before - This is the input files (referenced on top of each other).  The oblong boxes are the clippers, and the cells and text are being clipped.  It's hard to see at this zoom, but every spot is an X cell.

 

 

Microstation Before Close-Up  - This is a zoom in on the lower right of the Before image, showing that everything is a cell clearly.

 

 

Microstation After -  This is the result of the clip.  Red features are still cells (color changed for visibility), white dots are now active points.

 

 

Microstation After Close-Up  - This is a zoom on the same area as before, showing that the line of change seems to follow the bounding box of the clipping poly.

 

 

One thing I noted when setting this up is that the clip file has nested and intersecting polys in it (building lines and obscured lines).  I didn't think this would matter, but does it?
david_r
Rank
Userlevel 4
Hi,

 

 

thanks for the screenshots. I'm not sure what's going on and I'm in no way an expert on dgn cells, but:

 

  • Have you read the FAQ on FME and Microstation cells?
  • Could it be that your cells enter the Clipper as aggregates? Try inserting an AggregateFilter before the Clipper to see where your features are going. If so, you might try to insert a Deaggregator.
Also consider using feature inspection points in the workspace. It will enable you to pause the execution of the workspace and to inspect the features as they enter and exit the different transformers. It is very handy for debugging.

 

 

David
A
Badge +1
1.) I have read that, and I've worked with .dgn cells before.  I'm handling these cells in a way that has let cells pass through and be rewritten many times before.

 

 

2.) I don't think there's any way they could be aggregated, but I can't say for sure, so I'll try that ... didn't change anything.

 

 

3.) The dataflow here is Input -> Clipper -> Output, so there's no place to put an inspector.  If I inspect the output, it doesn't show what's really a cell and what isn't, and the attributes indicate that everything should be a cell.
david_r
Rank
Userlevel 4
Hi,

 

 

thanks for the info. Strange indeed. Consider sending a workspace with some sample data to Safe support so that they can take a look at it for you.

 

 

Regarding point 3), I was not thinking about an Inspector, but a feature inspection point. Take a look at the article I linked to -- it can be a fantastic help when debugging a workspace.

 

 

David

Reply


Community Stats

28,724
Posts
109,516
Replies
36,561
Members
Terms & ConditionsCookie settings
A product of
Safe Software respectfully acknowledges that we live, learn and work on the traditional and unceded territories of the Kwantlen, Katzie, and Semiahmoo First Nations.