ValueError raised when adding columns to a new Metadata instance

[pvk-developer]

Environment Details

Please indicate the following details about the environment in which you found the bug:

• SDV version: 1.17.0
• Python version: any
• Operating System: any

Error Description

When creating a new Metadata instance and trying to add new columns, a ValueError is raised indicating that it has more than one table.

Steps to reproduce

from sdv.metadata import Metadata

metadata = Metadata()
metadata.add_column(column_name='a', sdtype='numerical')

ValueError: Metadata contains more than one table, please specify the `table_name`.

Notice that the metadata is an empty Metadata object:

In [7]: metadata
Out[7]: 
{
    "tables": {},
    "relationships": [],
    "METADATA_SPEC_VERSION": "V1"
}

Expected Behavior

We should raise a proper error message:

The metadata object is empty, please use `Metadata.detect_from_dataframe` to populate the metadata.

[srinify]

Hmm IMO I'm not sure the error message is necessarily what we want to convey @pvk-developer because this implies people are only using the standard synthesizers. cc @npatki

With DayZ, you wouldn't be able to use detect_from_dataframe() or if you read in metadata from the BigQuery connector -- so this feels like too specific of an error message?

So I'm wondering if we should instead try to either nudge the user to create a table in the metadata (using the Metadata.create_table() method):

metadata = Metadata()
metadata.add_table('step_counts')