Glad to hear it's working OK now. The first tip I mentioned is an issue that all users should be aware of when using a transformer that transfers attributes from one feature to another.
The issue is that you can sometimes inadvertently transfer format attributes too. This could lead to your point features being tagged with the mapinfo_type attribute from the line features (for example) giving you a point feature being identified as a line. When you write the data it might then get thrown out as being structurally incorrect. The issue would only occur when... 1) Transferring attributes from one feature to another - eg NeighborFinder, FeatureMerger etc 2) Transferring between features with different geometries - eg point to a line 3) Reading and Writing to the same format (writing to a different format would invalidate the source format attributes so it wouldn't matter what they were) Just a tip y'all may find useful. Mark Mark Ireland, Product Specialist Safe Software Inc. Surrey, BC, CANADA [EMAIL PROTECTED] http://www.safe.com Solutions for Spatial Data Translation, Distribution and Access --- In [email protected], "latrobegis" <[EMAIL PROTECTED]> wrote: > > Thanks for the tips Mark and I have now sorted it out. > > Everything looked fine but there was a clue in the log as a couple > of candidate features didn't match and they all should have. It > turns out that there were some line segments that were not visible > and didn't have an ID and these had been replaced by new segments in > the same position, which explained why the distance attribute was > populated but the SegmentID was not. The visualiser didn't help in > this case as the points (base features) had all passed and matched > to these old segments in some cases. > > > > --- In [email protected], "mark2atsafe" <mark.ireland@> wrote: > > > > Hi, > > Difficult to diagnose this one, but here are some suggestions... > > > > 1) How are you checking the output? My advice would be to attach a > > Visualizer to the neighborFinder output port and see what it looks > > like there. There's a few unexpected effects of using > NeighborFinder > > that sometimes affect how data is output. If it looks OK in the > > Visualizer then check the format attributes and the log file to > make > > sure everything is getting written correctly. > > > > 2) Make sure the points and lines are correctly separated into BASE > > and CANDIDATES. Sorry if this is obvious but you need to make sure > > that all BASE features actually do have the ObjectID you're > looking for. > > > > 3) Apply a list to the neighborFinder. In your version the list > will > > show all lines whose bounding box the point falls within (a Tested > > Candidate list). If the expected line isn't in the list (and you > can > > prove it should be by creating the bounding box for it and > checking in > > the FME Viewer) then it's a problem you need to email to > support@ > > > > 4) If the line appears in the list then check it is actually within > > the correct distance. You can prove this graphically by using the > > Bufferer to create buffers of 'max_distance' size around the line. > Not > > that I'd disbelieve the _distance attribute, I just like to see > things > > with my own eyes. > > > > If you've still got problems then send it to support@ - it > > would be a real help if you can cut the datasets down to a single > > point and line feature that ought to match, but don't. That way > we'll > > be able to recreate and diagnose the problem a lot easier. > > > > Hope this helps, > > > > Mark > > > > Mark Ireland, Product Specialist > > Safe Software Inc. Surrey, BC, CANADA > > support@ http://www.safe.com > > Solutions for Spatial Data Translation, Distribution and Access > > > > > > --- In [email protected], "latrobegis" <DeanMo@> wrote: > > > > > > I am using Workbench Build 2182 and some MapInfo tab files > trying to > > > populate a set of 40,000 points with the ID from the closest > line > > > segment and not all the points are being populated. Of the > 40,000 > > > points there are about 500 that are not being given the > segmentID. > > > > > > Each of the points are within the range of the "Maximum > Distance" from > > > the NeighborFinder as the _distance attribute is being > populated. > > > Some line segments have only a few points populated while most > have > > > all points populated. However if I redo the workbench and I > test for > > > the worst line segment and then use the NeighborFinder it > populates > > > the points on that line segment fine. > > > > > > Any help would be appreciated thanks. > > > > > > Join us at the FME Worldwide User Conference Sept. 21-22, 2006 Vancouver BC Canada. For more information, visit www.safe.com/2006uc. Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/fme/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
