(Version 10.0 sp5)
The Check Geometry tool does dot work with an SDE feature class. Is there a tool or command that does the same thing I can use?
Geometry loaded in an enterprise geodatabase feature class has already been validated with the most rigorous validation available (the query stream will fail with a low-level transfer error if it was corrupted after loading)..
Vince- I have to export feature class to a Shapefile for use in an outside application. When I do, I get several empty geometry features which goof on me. I thought maybe if I compressed the SDE db that might help; in reality the export process created more empty geometries after compressing. (I've imported, exported, and projected to get the shapefile. All goof on me...)
Is this behavior indicative of anything I can correct?
Which database? How are you exporting the features? Does the table have CAD objects? What exact error is generated?
Tech Support may be necessary to discover what is going on.
The original gdb is in Sql Server 2008; Version 10.0 SDE, fully patched.
In addition to making a shapefile out of the feature class, it needs to be in a different projection than the original.
I have exported to a shapefile, and then reprojected that shapefile.
I have set the data frame properties to the desired output projection, added the original fc and exported using the data frame projection.
I have used the project tool to re-project the original fc to the desired output projection shapefile.
I have used the Feature Class to Feature Class tool, with the original projection and by setting the output environment to the desired projection.
I misspoke: you can't import to a shapefile. My bad.
After performing any of the above methods, when viewing the subsequent shapefile it renders a view indicative of projection issues; a small cluster of features in the upper right corner. I run the Check Geometry tool, and it lists the Empty Geometry shapes in the resulting dbf table. Then I simply relate the table to the feature class an delete the offending records. After that, the extent behaves as it should and I'm on my way...
So it sounds like the collapse is occurring during projection. You might just check for rows with very small areas.
It's a street polyline feature class; I looked for 0 (zero) length, but found nothing. When I exported to a fgdb or imported to a fgdb and ran the check geometry tool, the dbf warns of "short segments". Your comment about the projection has merit; I'm going from stateplane feet to utm meters. My guess is a sub-meter length segment becomes empty. However, I need to take a closer look at the original data; a segment of street that is less than 1 meter (39.3701 inches) is a problem unto itself.
The check geometry dbf is a handy table to work with; it provides a street name. I use that to select all the SDE features with the same street name. The ones that appear in the SDE data appear to be edits that have superceded the ones that error out. It seems to me that in the export/import/projection process captures these phantom records and chokes the shapefile.
I doubt versioning is broken, so there's something else going on. Tech Support may be able to help determine exactly what.
Retrieving data ...