join should act like upload

[burritojustice]

join currently expects a space ID that contains the geometries to be joined to a csv data table via virtual spaces -- it generates a space ID for the csv data table and a virtual space ID.

There is no way for join to update an existing data space to update the virtual space. While you can upload to the space, you need to specify --nocoords. This is also inconsistent with upload that either adds data to a specified space ID, or generates one if not present.

We should change join to accept a space ID in order to update the space containing the csv data table, and specify the geometry space ID using an argument

here xyz join -f my.csv --target space_with_geometries
or perhaps
here xyz join -f my.csv --geom space_with_geometries -i important_property

If a space ID is present, this could either be the space ID of the virtual space, or the space ID of the data table (the second space in the virtual space argument). Need to discuss.

here xyz join spaceID -f data_update.csv

• how do we maintain consistency with -i -- possibly record column used as -i / feature ID it in theuser block of the space object #228 ?
• what about --groupby, --keys -- maybe reference upload command history?

we also need to consider the interactive mode for join