I have a line shapefile with 100+ fields and a separate dbf file with 100+ fields. I'd like to programatically (using arcpy and Python 3) eliminate all but 3 fields in both shapefile and the dbf, and then join on a common field. The desired output would only contain 5 fields.

If I do a simple join:

joined = arcpy.AddJoin(path_to_shp, "id_field", path_to_dbf, "id_field")

I cannot delete the extra fields. And if I write this joined table to a file, the joined field names become unrecognizable due to length constraints.

I was looking at the 'field_info' parameter of MakeFeatureLayer, but I can't find any examples of how this would be implemented, and it seems like none of the tools I'd normally use for this take FieldMappings.

  • you mention arcpy, but would geopandas be an option? the geopandas-based solution for this problem is quite simple. – Paul H Mar 20 at 21:56

AddJoin is an in memory join which typically you make and break within a session. Join Field tool does a permanent join of data.

But you are talking about joining 100's of fields to 100's of fields which may exceed the design limitations of a shapefile.

A better and more robust solution is to hide the fields you don't want and then use the Copy Features tool to create a temporary copy, this will honour hidden fields and the resulting output would be a dataset of the 3 fields. Copy Rows should honour hidden fields too. Then use Join Field to bring the two datasets together.


In the off chance you or a future reader has the option of doing this geopandas, here's the basic approach you'd take:

  1. Read both files with geopandas.read_file
  2. select only the columns you want to keep + the 'geometry' column of the shapefile
  3. use the .merge method
  4. pipe the merge result to geopandas.GeoDataFrame in case it returned a dataframe and not a geodataframe

Here's what that would look like:

import geopandas

new_gdf = (
        .loc[:, ['A', 'B', 'C', 'geometry']]
            geopandas.read_file('/path/to/meta.dbf').loc[:, ['A', 'D', 'E']],
        .pipe(geopandas.GeoDataFrame, geometry='geometry')


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.