Merging Personal Geodatabase (MDB) data into ArcSDE using FME?

Question

Im a novice in FME and I have a MDB file that i want to write/insert to an SDE database using FME. I created a simple translation that generally reads the data from the MDB file and writes it to the SDE database, although the translation is successful and it prints out the number of written rows, when i checked the SDE database, it wasn't able to right any data on it.

i then checked the log and i found this specific warning:

MS SQL Server (Spatial) Writer: Attribute data ' too large to fit in specified data type for columnSHAPE'

The shape column holds the geometry data in the MDB file. FME sets the SHAPE datatype to 'oleobject' in the Reader and I noticed that the SHAPE column in the writer was automatically converted to 'text' datatype.

Is there anything else i need to configure to make this work?

UPDATE: I am using FME Database Edition 2014 (20140120 - Build 14235 - WIN64)

Here are some screenshots of my FME translation:

Answer 1

FME's MDB Reader is only a reader for pure tabular data (without geometry). It interpretes the shape column as some kind of binary data, but doesn't know that it contains geometry actually. Therefore the content then neither doesn't fit into the datatype FME initially uses for that nor does it know how to store any geometry into the SDE layer (again, because it haven't read any geometry).

If the MDB File is a ESRI personal Geodatabase use the ESRI Geodatabase reader instead.

Use the format Identifier GEODATABASE_MDB to define, that it is a Geodatabase stored in a mdb file, as there is also the option to store it in a RDBMS (enterprise geodatabase based on SDE) or in a directory (File Geodatabase).

If you use it correctly, you will see, that FME never ever presents the column containing the geometry as a mappable field (as "shape" with the MDB reader) because it only presents attribute fields. Geometries are handled in the background.

When you have defined your SDE writer, you can right click your Geodatabase reader feature type and select "duplicate on writer" to define a 1:1 copy of your feature type schema in SDE.

In the writer feature types properties you can select how the feature class should be updated, i.e.:

• Should it drop and recreate the feature class with every run of the workbench
• Should it only truncate the table and append the features then
• Should it only append

With "Bulk insert" set to Yes (see screenshot) you can speed up writing up from 6 to 10 times.There aren't any situations where this could really cause a problem, except when you want to update small sets of data and can't proof the validity before appending them to your destination table.

Screenshot via @Mapperz using FME 2014 (SP1 build 14287)