BUG: use original schema when appending (#318)

* BUG: use original schema when appending

Don't overwrite table schema when appending to an existing table

* python 3.5 doesn't support f-string

* cln: refactor `to_gbq` to avoid unnecessary extra table GET HTTP calls

pandas-gbq already gets the table metadata when checking if a table
exists. This refactoring avoids extra calls to get the table metadata
when checking the schema.

also, fix a bug where update_schema appends columns that aren't in the
dataframe to the schema sent in the API request

* doc: add fix to changelog

* doc: revert accidental whitespace change

Co-authored-by: Tim Swast <[email protected]>

.github/PULL_REQUEST_TEMPLATE.md

@@ -0,0 +1,4 @@
+- [ ] closes #xxxx
+- [ ] tests added / passed
+- [ ] passes `nox -s blacken lint`
+- [ ] `docs/source/changelog.rst` entry

docs/source/changelog.rst

@@ -1,6 +1,14 @@
 Changelog
 =========
 
+.. _changelog-0.13.2:
+
+0.13.2 / TBD
+------------
+
+- Fix ``Provided Schema does not match Table`` error when the existing table
+  contains required fields. (:issue:`315`)
+
 .. _changelog-0.13.1:
 
 0.13.1 / 2020-02-13

pandas_gbq/gbq.py

@@ -5,6 +5,15 @@
 
 import numpy as np
 
+# Required dependencies, but treat as optional so that _test_google_api_imports
+# can provide a better error message.
+try:
+    from google.api_core import exceptions as google_exceptions
+    from google.cloud import bigquery
+except ImportError:  # pragma: NO COVER
+    bigquery = None
+    google_exceptions = None
+
 try:
     # The BigQuery Storage API client is an optional dependency. It is only
     # required when use_bqstorage_api=True.
@@ -388,7 +397,6 @@ def sizeof_fmt(num, suffix="B"):
         return fmt % (num, "Y", suffix)
 
     def get_client(self):
-        from google.cloud import bigquery
         import pandas
 
         try:
@@ -429,7 +437,6 @@ def run_query(
     ):
         from concurrent.futures import TimeoutError
         from google.auth.exceptions import RefreshError
-        from google.cloud import bigquery
 
         job_config = {
             "query": {
@@ -640,15 +647,6 @@ def schema(self, dataset_id, table_id):
         except self.http_error as ex:
             self.process_http_error(ex)
 
-    def _clean_schema_fields(self, fields):
-        """Return a sanitized version of the schema for comparisons."""
-        fields_sorted = sorted(fields, key=lambda field: field["name"])
-        # Ignore mode and description when comparing schemas.
-        return [
-            {"name": field["name"], "type": field["type"]}
-            for field in fields_sorted
-        ]
-
     def verify_schema(self, dataset_id, table_id, schema):
         """Indicate whether schemas match exactly
 
@@ -672,43 +670,12 @@ def verify_schema(self, dataset_id, table_id, schema):
             Whether the schemas match
         """
 
-        fields_remote = self._clean_schema_fields(
+        fields_remote = pandas_gbq.schema._clean_schema_fields(
             self.schema(dataset_id, table_id)
         )
-        fields_local = self._clean_schema_fields(schema["fields"])
-
+        fields_local = pandas_gbq.schema._clean_schema_fields(schema["fields"])
         return fields_remote == fields_local
 
-    def schema_is_subset(self, dataset_id, table_id, schema):
-        """Indicate whether the schema to be uploaded is a subset
-
-        Compare the BigQuery table identified in the parameters with
-        the schema passed in and indicate whether a subset of the fields in
-        the former are present in the latter. Order is not considered.
-
-        Parameters
-        ----------
-        dataset_id : str
-            Name of the BigQuery dataset for the table
-        table_id : str
-            Name of the BigQuery table
-        schema : list(dict)
-            Schema for comparison. Each item should have
-            a 'name' and a 'type'
-
-        Returns
-        -------
-        bool
-            Whether the passed schema is a subset
-        """
-
-        fields_remote = self._clean_schema_fields(
-            self.schema(dataset_id, table_id)
-        )
-        fields_local = self._clean_schema_fields(schema["fields"])
-
-        return all(field in fields_remote for field in fields_local)
-
     def delete_and_recreate_table(self, dataset_id, table_id, table_schema):
         table = _Table(
             self.project_id, dataset_id, credentials=self.credentials
@@ -1141,7 +1108,6 @@ def to_gbq(
     """
 
     _test_google_api_imports()
-    from pandas_gbq import schema
 
     if verbose is not None and SHOW_VERBOSE_DEPRECATION:
         warnings.warn(
@@ -1168,25 +1134,31 @@ def to_gbq(
         credentials=credentials,
         private_key=private_key,
     )
+    bqclient = connector.client
     dataset_id, table_id = destination_table.rsplit(".", 1)
 
-    table = _Table(
-        project_id,
-        dataset_id,
-        location=location,
-        credentials=connector.credentials,
-    )
-
     default_schema = _generate_bq_schema(dataframe)
     if not table_schema:
         table_schema = default_schema
     else:
-        table_schema = schema.update_schema(
+        table_schema = pandas_gbq.schema.update_schema(
             default_schema, dict(fields=table_schema)
         )
 
     # If table exists, check if_exists parameter
-    if table.exists(table_id):
+    try:
+        table = bqclient.get_table(destination_table)
+    except google_exceptions.NotFound:
+        table_connector = _Table(
+            project_id,
+            dataset_id,
+            location=location,
+            credentials=connector.credentials,
+        )
+        table_connector.create(table_id, table_schema)
+    else:
+        original_schema = pandas_gbq.schema.to_pandas_gbq(table.schema)
+
         if if_exists == "fail":
             raise TableCreationError(
                 "Could not create the table because it "
@@ -1199,16 +1171,20 @@ def to_gbq(
                 dataset_id, table_id, table_schema
             )
         elif if_exists == "append":
-            if not connector.schema_is_subset(
-                dataset_id, table_id, table_schema
+            if not pandas_gbq.schema.schema_is_subset(
+                original_schema, table_schema
             ):
                 raise InvalidSchema(
                     "Please verify that the structure and "
                     "data types in the DataFrame match the "
                     "schema of the destination table."
                 )
-    else:
-        table.create(table_id, table_schema)
+
+            # Update the local `table_schema` so mode matches.
+            # See: https://github.com/pydata/pandas-gbq/issues/315
+            table_schema = pandas_gbq.schema.update_schema(
+                table_schema, original_schema
+            )
 
     if dataframe.empty:
         # Create the table (if needed), but don't try to run a load job with an

pandas_gbq/schema.py

@@ -3,6 +3,59 @@
 import copy
 
 
+def to_pandas_gbq(client_schema):
+    """Given a sequence of :class:`google.cloud.bigquery.schema.SchemaField`,
+    return a schema in pandas-gbq API format.
+    """
+    remote_fields = [
+        field_remote.to_api_repr() for field_remote in client_schema
+    ]
+    for field in remote_fields:
+        field["type"] = field["type"].upper()
+        field["mode"] = field["mode"].upper()
+
+    return {"fields": remote_fields}
+
+
+def _clean_schema_fields(fields):
+    """Return a sanitized version of the schema for comparisons.
+
+    The ``mode`` and ``description`` properties areis ignored because they
+    are not generated by func:`pandas_gbq.schema.generate_bq_schema`.
+    """
+    fields_sorted = sorted(fields, key=lambda field: field["name"])
+    return [
+        {"name": field["name"], "type": field["type"]}
+        for field in fields_sorted
+    ]
+
+
+def schema_is_subset(schema_remote, schema_local):
+    """Indicate whether the schema to be uploaded is a subset
+
+    Compare the BigQuery table identified in the parameters with
+    the schema passed in and indicate whether a subset of the fields in
+    the former are present in the latter. Order is not considered.
+
+    Parameters
+    ----------
+    schema_remote : dict
+        Schema for comparison. Each item of ``fields`` should have a 'name'
+        and a 'type'
+    schema_local : dict
+        Schema for comparison. Each item of ``fields`` should have a 'name'
+        and a 'type'
+
+    Returns
+    -------
+    bool
+        Whether the passed schema is a subset
+    """
+    fields_remote = _clean_schema_fields(schema_remote.get("fields", []))
+    fields_local = _clean_schema_fields(schema_local.get("fields", []))
+    return all(field in fields_remote for field in fields_local)
+
+
 def generate_bq_schema(dataframe, default_type="STRING"):
     """Given a passed dataframe, generate the associated Google BigQuery schema.
 
@@ -59,9 +112,6 @@ def update_schema(schema_old, schema_new):
         if name in field_indices:
             # replace old field with new field of same name
             output_fields[field_indices[name]] = field
-        else:
-            # add new field
-            output_fields.append(field)
 
     return {"fields": output_fields}
 

tests/system/test_gbq.py

@@ -12,6 +12,7 @@
 import pytz
 
 from pandas_gbq import gbq
+import pandas_gbq.schema
 
 
 TABLE_ID = "new_test"
@@ -1637,7 +1638,7 @@ def test_retrieve_schema(gbq_table, gbq_connector):
     }
 
     gbq_table.create(table_id, test_schema)
-    actual = gbq_connector._clean_schema_fields(
+    actual = pandas_gbq.schema._clean_schema_fields(
         gbq_connector.schema(gbq_table.dataset_id, table_id)
     )
     expected = [
@@ -1649,48 +1650,39 @@ def test_retrieve_schema(gbq_table, gbq_connector):
     assert expected == actual, "Expected schema used to create table"
 
 
-def test_schema_is_subset_passes_if_subset(gbq_table, gbq_connector):
-    # Issue #24 schema_is_subset indicates whether the schema of the
-    # dataframe is a subset of the schema of the bigquery table
-    table_id = "test_schema_is_subset_passes_if_subset"
+def test_to_gbq_does_not_override_mode(gbq_table, gbq_connector):
+    # See: https://github.com/pydata/pandas-gbq/issues/315
+    table_id = "test_to_gbq_does_not_override_mode"
     table_schema = {
         "fields": [
-            {"name": "A", "type": "FLOAT"},
-            {"name": "B", "type": "FLOAT"},
-            {"name": "C", "type": "STRING"},
-        ]
-    }
-    tested_schema = {
-        "fields": [
-            {"name": "A", "type": "FLOAT"},
-            {"name": "B", "type": "FLOAT"},
+            {
+                "mode": "REQUIRED",
+                "name": "A",
+                "type": "FLOAT",
+                "description": "A",
+            },
+            {
+                "mode": "NULLABLE",
+                "name": "B",
+                "type": "FLOAT",
+                "description": "B",
+            },
+            {
+                "mode": "NULLABLE",
+                "name": "C",
+                "type": "STRING",
+                "description": "C",
+            },
         ]
     }
 
     gbq_table.create(table_id, table_schema)
-    assert gbq_connector.schema_is_subset(
-        gbq_table.dataset_id, table_id, tested_schema
+    gbq.to_gbq(
+        pandas.DataFrame({"A": [1.0], "B": [2.0], "C": ["a"]}),
+        "{0}.{1}".format(gbq_table.dataset_id, table_id),
+        project_id=gbq_connector.project_id,
+        if_exists="append",
     )
 
-
-def test_schema_is_subset_fails_if_not_subset(gbq_table, gbq_connector):
-    # For pull request #24
-    table_id = "test_schema_is_subset_fails_if_not_subset"
-    table_schema = {
-        "fields": [
-            {"name": "A", "type": "FLOAT"},
-            {"name": "B", "type": "FLOAT"},
-            {"name": "C", "type": "STRING"},
-        ]
-    }
-    tested_schema = {
-        "fields": [
-            {"name": "A", "type": "FLOAT"},
-            {"name": "C", "type": "FLOAT"},
-        ]
-    }
-
-    gbq_table.create(table_id, table_schema)
-    assert not gbq_connector.schema_is_subset(
-        gbq_table.dataset_id, table_id, tested_schema
-    )
+    actual = gbq_connector.schema(gbq_table.dataset_id, table_id)
+    assert table_schema["fields"] == actual