sdv.metadata.dataset.Metadata.add_table

Metadata.add_table(name, data=None, fields=None, fields_metadata=None, primary_key=None, parent=None, foreign_key=None, constraints=None)[source]

Add a new table to this metadata.

fields list can be a mixture of field names, which will be build automatically from the data, or dictionaries specifying the field details. If a field needs to be analyzed, data has to be also passed.

If parent is given, a relationship will be established between this table and the specified parent.

Parameters
  • name (str) – Name of the new table.

  • data (str or pandas.DataFrame) – Table to be analyzed or path to the csv file. If it’s a relative path, use root_path to find the file. Only used if fields is not None. Defaults to None.

  • fields (list) – List of field names to build. If None is given, all the fields found in the data will be used. Defaults to None.

  • fields_metadata (dict) – Metadata to be used when creating fields. This will overwrite the metadata built from the fields found in data. Defaults to None.

  • primary_key (str) – Field name to add as primary key, it must not exists. Defaults to None.

  • parent (str) – Table name to refere a foreign key field. Defaults to None.

  • foreign_key (str) – Foreign key field name to parent table primary key. Defaults to None.

  • constraints (list[Constraint, dict]) – List of Constraint objects or dicts representing the constraints for the given table.

Raises

ValueError – If the table name already exists or data is not passed and fields need to be built from it.