I have a large .gdb of all blocks in the United States, available from Census. I would like to create a point shape file of all census tract centroids in the US.

I have QGIS with FileGDB enabled, as well as R with rgdal, so reading the data is not a technical problem. My issue is that reading the layer into R has taken 2 hours and used up 56 Gb (I have 64) of memory, and it's not done. I know exactly what I want to do:

  1. Create a tract field by pulling the first 7 characters out of the GEOID field
  2. Dissolve the blocks on this field
  3. Calculate the centroids of the new geographies
  4. Export as a points shape file, or even as another layer in the database.

QGIS shows the data, but the field calculations take just about as long. It seems that I should be able to send these commands into the database rather than load the database and try to do it interactively. I might need some hand holding to do this, or at least good pointers to the right manuals.

On the other hand, is there a way to pull the database out in chunks? I am running Ubuntu 14.04 on a 12-core machine, so parallelization is always a good option...

  • 1
    File geodatabase is not an RDBMS. It's just a file format, so there is nothing to send commands "into".
    – Vince
    Commented Aug 19, 2014 at 15:26
  • Ah, this is good to know. Commented Aug 19, 2014 at 15:30


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.