Unable to validate just 1 table of a multi-table schema (#2679)

Author: frances-h

sdv/metadata/metadata.py

@@ -297,7 +297,8 @@ def validate_table(self, data, table_name=None):
                 'to validate.'
             )
 
-        return self.validate_data({table_name: data})
+        self._validate_table_exists(table_name)
+        return self._validate_data({table_name: data}, table_name)
 
     def get_column_names(self, table_name=None, **kwargs):
         """Return a list of column names that match the given metadata keyword arguments."""

sdv/metadata/multi_table.py

@@ -914,17 +914,20 @@ def _validate_foreign_keys(self, data):
 
         return [error_msg] if error_msg else []
 
-    def validate_data(self, data):
-        """Validate the data matches the metadata.
+    def _validate_data(self, data, table_name=None):
+        """Validate the given data matches the metadata.
 
         Checks the following rules:
-            * all tables of the metadata are present in the data
             * every table of the data satisfies its own metadata
-            * all foreign keys belong to a primay key
+            * if no table_name provided, all tables of the metadata are present in the data
+            * if no table_name provided, that all foreign keys belong to a primay key
 
         Args:
             data (dict):
                 A dictionary of table names to pd.DataFrames.
+            table_name (str, optional):
+                The specific table to validate. If set, only validates the data for the
+                table. If None, validates the data for all tables. Defaults to None.
 
         Raises:
             InvalidDataError:
@@ -938,13 +941,35 @@ def validate_data(self, data):
             raise InvalidMetadataError('Please pass in a dictionary mapping tables to dataframes.')
 
         errors = []
-        errors += self._validate_missing_tables(data)
+        errors += self._validate_missing_tables(data) if not table_name else []
         errors += self._validate_all_tables(data)
-        errors += self._validate_foreign_keys(data)
+        errors += self._validate_foreign_keys(data) if not table_name else []
 
         if errors:
             raise InvalidDataError(errors)
 
+    def validate_data(self, data):
+        """Validate the data matches the metadata.
+
+        Checks the following rules:
+            * every table of the data satisfies its own metadata
+            * all tables of the metadata are present in the data
+            * all foreign keys belong to a primay key
+
+        Args:
+            data (dict):
+                A dictionary of table names to pd.DataFrames.
+
+        Raises:
+            InvalidDataError:
+                This error is being raised if the data is not matching its sdtype requirements.
+
+        Warns:
+            A warning is being raised if ``datetime_format`` is missing from a column represented
+            as ``object`` in the dataframe and its sdtype is ``datetime``.
+        """
+        self._validate_data(data)
+
     def add_table(self, table_name):
         """Add a table to the metadata.
 

tests/unit/metadata/test_metadata.py

@@ -4,7 +4,6 @@
 import pandas as pd
 import pytest
 
-from sdv.errors import InvalidDataError
 from sdv.metadata.errors import InvalidMetadataError
 from sdv.metadata.metadata import Metadata
 from sdv.metadata.single_table import SingleTableMetadata
@@ -584,18 +583,15 @@ def test_validate_table(self):
         )
         table = get_multi_table_data()['nesreca']
 
-        expected_error_wrong_name = re.escape(
-            'The provided data does not match the metadata:\n'
-            "The provided data is missing the tables {'nesreca'}."
-        )
+        expected_error_wrong_name = re.escape("Unknown table name ('wrong_name').")
         expected_error_mutli_table = re.escape(
             'Metadata contains more than one table, please specify the `table_name` to validate.'
         )
 
         # Run and Assert
         metadata_single_table.validate_table(table)
         metadata_single_table.validate_table(table, 'nesreca')
-        with pytest.raises(InvalidDataError, match=expected_error_wrong_name):
+        with pytest.raises(InvalidMetadataError, match=expected_error_wrong_name):
             metadata_single_table.validate_table(table, 'wrong_name')
         with pytest.raises(InvalidMetadataError, match=expected_error_mutli_table):
             metadata_multi_table.validate_table(table)