shrug-l: Data Drift in GeoDataBases

[Arbogast, Rene]

  I would guess that there are mathematical calculations in the =
conversion
process which are 'almost' guaranteed to generate some mathematical =
rounding
errors. =20

Rene'

-----Original Message-----
From: Mark.Welsh@dot.state.fl.us [mailto:Mark.Welsh@dot.state.fl.us]
Sent: Thursday, October 17, 2002 6:18 PM
To: shrug-l@lists.dep.state.fl.us
Subject: shrug-l: Data Drift in GeoDataBases


Hello,

I have noticed than when I import a coverage or shape file into a =
personal
geodatabase (GDB), the new feature class in the GDB exhibits spatial =
drift.
While the movement of the data is slight (1/1000 of a meter for our road
network), I was shocked to see that when I zoomed in real tight and
compared the GDB feature class to the original data, there was an =
offset.
I then decided to use a spatial select where I picked an arc from the =
GDB
and asked ArcMap to select the same arc (contained by) from the coverage
and no selection was made.  So on the one hand, the problem seems small
(0.001 meter difference) but on the other hand, the problem is =
significant
when you try certain types of spatial selections.  This problem does not
manifest itself when I take the coverage, convert it to a shapefile, and
try to select an arc from the shape file by using the coverage as the
selector theme (in both ArcMap and ArcView 3.1) so the problem seems to =
be
with the GDB.  Also, when I convert the GDB feature class back out to a
shapefile or coverage, the drift is now present in the converted data!

I have written VB code to bring all sorts of spatial data into a GDB =
using
the IFeatureDataConverter.  Its function call includes an =
IGeometryDefEdit
interface which allows you to manually set things like precision, x/y
domain extent, and scale.  However, when I doe this via programming, or
manually via arccatalog import tools, I can only get the data to be a
little "closer" to the original, but never exact!

I talked to some geodatabase product development team members who (a) =
could
not provide me with an answer to my request "please tell me how to make =
the
GDB data the same (exactly) as the source coverage or shape file and (b)
they told me I was too worried about such a minor difference between =
source
and original data.

Has anybody else experienced/noticed this problem?  If so, have you =
found a
way to fix this via code or otherwise?  Do people think that I'm being =
to
picky about converted data being "different" from source?  Help.......

Thanks!

Mark Welsh
GIS Applications Coordinator
Transportation Statistics Office
Florida Department of Transportation
(850) 414-4722
mark.welsh@dot.state.fl.us

_______________________________________________
SHRUG-L mailing list
SHRUG-L@lists.dep.state.fl.us
http://lists.dep.state.fl.us/mailman/listinfo/shrug-l