Exporting After Join

[KevinLaughlin]

This is frustrating because this is something I've done a thousand times before. However, I have a feature class called Parcels and I'm joining it to an assessment table based on a common field called parcel number. In ArcMap, it works fine. When I export the table, or export a subset selection, the table columns are all <null> in the exported FC. It does the same thing when I create a layer based off of a selection. It exports field names, but not the data in them.

I've recently just upgraded to 10.3, so I'm not sure if this is a bug therefrom, or I'm just missing something.

[SusanJones · Accepted solution]

Hi Kevin

sounds like the source is not a Geodatabase Table or recognised format.

It might be easiest if you import the table into a file geodatabase. The objectids will be added automatically.

Susan

[KevinLaughlin]

Susan Jones‌

I took your advice and it worked. I was going to try one last thing before I went down the hall to try it in another office that hadn't updated from 10.2.

I imported the table into a file geodatabase. Performed the join with the imported table, and exported it and it worked properly.

Thanks,

Kevin

[KevinLaughlin]

Kinda/sorta. I created an empty table and copied the original to it, which apparently wasn't the same thing.

This time, I simply created and empty file geodatabase and imported the original table (without OIDs) and it worked.

I'm confused myself.

[KimOllivier]

What a wonderful insight to this problem that has bugged me for years. I avoid all exported joins because of the nulls, and I could never work out why. All I now have to do is check for the existence of a FID! My workaround is to create a python dictionary and run a cursor which is actually faster.

Might I also suggest SCHEMA.INI on your text files before you import them with TabletoTable? This is a standard Microsoft function that is used by ArcGIS to add a correct database schema to a CSV or TXT file. This avoids the dreaded 255 char wide fields, and properly defines integers instead of floats. All this will make joins more successful when it is finally in a geodatabase with a proper ObjectID.

[SusanJones]

Agreed Kim Ollivier The da (data access) module for arcpy has been revolutionary for Geoprocessing at in ArcGIS.

Rows are treated as tuples with these new cursors. Geoprocessing flies.