Topology errors?

[ChrisVolz]

My problem may be specific and difficult to answer. Perhaps someone has an idea how to deal with it...

I have a polygon dataset in an Esri GDB. This GDB can be exported to Interlis1 (.itf) with success. The validation afterwards, however, fails due to topology errors (overlaps, gaps,...).

My problem is: I cannot detect any errors in the feature dataset. Visually, everything is ok. The ArcMap-Topolgy-Check is ok. Importing the dataset in GRASS, cleaning the topogly, exporting to shp and re-importing in the GDB does not help. Using the Interlis-Viewer no problems can be visually detected.

It seems that the feature dataset is ok, but the validator does not succeed. Somebody has an idea what can cause the problem and how to tackle it? Has someone used the ili2fgdb-tool on non-trivial polygon-datasets with success?

best regards, Chris

[claeis]

• Try the additional option --trace, that gives a more detailed log.
• Try to isolate the curve segments, that reproduce the problem (example https://github.com/claeis/iox-ili/blob/master/src/test/data/Itf/Test2LV95_05AreaHole.itf)

[ChrisVolz]

Dear claeis

I have now isolated the problem further and I am quite sure, that something is wrong with the provided tools. My workflow:

1) I have downloaded an .itf file and .ili file from the Kanton Solothurn. 2) This .itf file and .ili file passes the validator. 3) I create a GDB from the .ili file (--schemaimport) using the ili2fgdb-tool 4) I import the .itf file (--import) to the created GDB using the ili2fgdb-tool 5) I export the GDB (--export) to a new .itf file using the ili2fgdb-tool 6) The validation of the exported .itf file fails.

In my opinion this must not happen. The data seems to be changed somewhere in this process. Unfortunately I have no idea how to tackle this problem...

regards

PS:

• In this conversion process, duplicated lines (or lines having nearly(?) equal coordinates) are generated.

• The imported data in the database seems correct. Probably the error is introduced during data export.

nat_dornach.zip

[ChrisVolz]

Additional note:

I tried the same workflow using POSTGIS instead of GDB. The same result - the exported .itf fails the validator...

[claeis]

which versions did you use?

[ChrisVolz]

my used versions:

• ili2fgdb-3.10.9
• ili2pg-3.11.0
• ilivalidator-1.5.0

ArcGIS 10.1

Postgresql 10.1, POSTGIS 2.4

[ChrisVolz]

Sorry for posting such a lot ;)I have further isolated the problem to following situation:

• I have a GDB with only 1 dataset containing 2 Polygons
• The 2 Polygons are topological correct
• I use ili2fgdb to export the GDB to .itf (java -jar ili2fgdb.jar --export --models Naturgefahrenkarte_SO_V11 --dbfile bbb.gdb export.itf)
• The export.tif file has "double lines" and, hence, is topological incorrect

The files (GDB and .ili) are attached to this posting. I hope this allows for easier bug-testing? doppelte_linien.zip

best regards

[ChrisVolz]

An additional info: The ili2fgdb-tools seem to work with INTERLIS2. The validation suceeds. The problem with duplicated lines is limited to INTERLIS1. Is there a tool for converting from INTERLIS2 to INTERLIS1 ?

Using the parameter -skipPolygonBuilding in the ilivalidator also eliminates the problem. It must be a problem of building polygons in INTERLIS1.