When I call resource.to_pandas() on a resource that has a primaryKey defined that key is used as an index for the dataframe. This is a reasonable default behavior, but took me some time to realize/debug.

Possible Solutions

1. Add an option for the to_pandas function of index=false, (with a default of true) which follows a similar pattern to pandas to_csv function.
2. Add additional documentation on the to_pandas function. The only thing I could find online for this was here: https://framework.frictionlessdata.io/docs/formats/pandas.html
   Adding additional documentation here that describes how primaryKey fields are used to index the dataframe would've cleared this up for me.

Current Workaround
Now that I know that to_pandas works this way I can call df.reset_index() [docs] to turn those index columns into regular columns, which is what I want in my situation.