In most cases, EDA is done at the REPL, where defdf sets up a package with the same name as the data-frame and interns the keys and properties. However when building a data pipeline for automated processing, this is unlikely to be the case. When testing a data pipeline it is useful to be able to use (describe <data-frame>) for debugging. The symbol handling for this use case, calling describe on a data-frame not created with defdf needs to account for this.