How can I split large File Geodatabase(FGDB) in Arcgis, It contains many Feature datasets and Feature classes. I want to split it without missing any feature or attribute and after some work I could be able to merge it into one FGDB. Purpose of splitting is, I have one FGDB and 8 staffs.If I split it into 4 parts it would be helpful to finish job earlier.

Say I have 1 GDB,10 datasets and 200 Feature classes.. GDB 1 should contain 10 datasets and 50 feature classes GDB 2 should contain 10 datasets and 50 feature classes like that.

Feature datasets should remain same because all 10 datasets contains all type of feature classes.

  • How are you wanting to split them? Does each new file geodatabase get ca 25% of the feature classes, ca 25% of the feature datasets, or maybe all feature classes and feature datasets but only ca 25% of the features in them? Do you plan to split any features or do you have a business rule that only whole features reach any new file geodatabase? If you are splitting by geography how are you planning to split standalone tables that may need to be joined to those features? Do you have any other data types/composites like rasters, topologies, networks, etc.
    – PolyGeo
    Aug 7 '17 at 4:43
  • Do you have SDE? If not you can clip the database, feature class by feature class, using a grid (2 x 2)... but trust me when I say that joining the database back together is not as simple as you might think. Aug 7 '17 at 4:43
  • If your staff is all on the same network, you should use a versioned database. Then you can all edit at the same time.
    – jbalk
    Aug 7 '17 at 4:49
  • 1
    SDE is Esri Spatial Database Engine, a technology that allows connection between ArcGIS and an enterprise database (like Oracle or PostgreSQL). SDE allows multi-user editing and check-out versioning. This is not the same as SDE personal that was distributed with v9.something. If you have ArcGIS Server Enterprise (standard or advanced) you should have also a SDE license; this would make your problem much simpler, or more correctly, redundant. Aug 7 '17 at 5:07
  • 1
    Documentation about versioning: desktop.arcgis.com/en/arcmap/10.3/manage-data/geodatabases/what-is-a-version.htm
    – jbalk
    Aug 7 '17 at 5:34

To accomplish what you describe in your clarified question you could:

  1. Make four copies of your file geodatabase (and keep the original so that you have an online backup to go to, in case your proof of concept proves otherwise).
  2. Decide which quarter of the feature classes are going to stay in each of the file geodatabases, and delete the other three-quarters of the feature classes from each.

This solution makes no suggestions for dealing with standalone tables or any other data types/composites like rasters, topologies, networks, etc.

To return your edited data to a single merged file geodatabase you could:

  1. copy one file geodatabase to a new "to be merged" file geodatabase; and
  2. copy in the feature classes (to the correct feature dataset) from the other three file geodatabases
  • Feature class by feature class is easier to join than area by area. No matter how careful your capture users are there will always be spatial and attribute mismatches at the joins if you're editing on an area by area basis. The trick is trying to split the work evenly between the datasets, it's not just a count of feature classes. Aug 7 '17 at 5:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.