refactor(api): remove schema (#10149)

BREAKING CHANGE: Removed hierarchical usage of schema. 

Ibis uses the following naming conventions:
- schema: a mapping of column names to datatypes
- database: a collection of tables
- catalog: a collection of databases

---------

Co-authored-by: Gil Forsyth <gforsyth@users.noreply.github.com>
Co-authored-by: Gil Forsyth <gil@forsyth.dev>

docs/backends/_utils.py

@@ -40,10 +40,7 @@ def find_member_with_docstring(member):
             if base not in resolved_bases:
                 resolved_bases.append(base)
 
-    # Remove `CanCreateSchema` and `CanListSchema` since they are deprecated
-    # and we don't want to document their existence.
-    filtered_bases = filter(lambda x: "schema" not in x.name.lower(), resolved_bases)
-    for base in filtered_bases:
+    for base in resolved_bases:
         try:
             parent_member = get_callable(base, member.name)
         except KeyError:

ibis/backends/__init__.py

@@ -741,44 +741,6 @@ def drop_database(
         """
 
 
-# TODO: remove this for 10.0
-class CanListSchema:
-    @util.deprecated(
-        instead="Use `list_databases` instead`", as_of="9.0", removed_in="10.0"
-    )
-    def list_schemas(
-        self, like: str | None = None, database: str | None = None
-    ) -> list[str]:
-        return self.list_databases(like=like, catalog=database)
-
-    @property
-    @util.deprecated(
-        instead="Use `Backend.current_database` instead.",
-        as_of="9.0",
-        removed_in="10.0",
-    )
-    def current_schema(self) -> str:
-        return self.current_database
-
-
-class CanCreateSchema(CanListSchema):
-    @util.deprecated(
-        instead="Use `create_database` instead", as_of="9.0", removed_in="10.0"
-    )
-    def create_schema(
-        self, name: str, database: str | None = None, force: bool = False
-    ) -> None:
-        self.create_database(name=name, catalog=database, force=force)
-
-    @util.deprecated(
-        instead="Use `drop_database` instead", as_of="9.0", removed_in="10.0"
-    )
-    def drop_schema(
-        self, name: str, database: str | None = None, force: bool = False
-    ) -> None:
-        self.drop_database(name=name, catalog=database, force=force)
-
-
 class CacheEntry(NamedTuple):
     orig_op: ops.Relation
     cached_op_ref: weakref.ref[ops.Relation]

ibis/backends/bigquery/__init__.py

@@ -25,7 +25,7 @@
 import ibis.expr.schema as sch
 import ibis.expr.types as ir
 from ibis import util
-from ibis.backends import CanCreateDatabase, CanCreateSchema
+from ibis.backends import CanCreateDatabase
 from ibis.backends.bigquery.client import (
     bigquery_param,
     parse_project_and_dataset,
@@ -155,7 +155,7 @@ def _postprocess_arrow(
     return table_or_batch.rename_columns(names)
 
 
-class Backend(SQLBackend, CanCreateDatabase, CanCreateSchema):
+class Backend(SQLBackend, CanCreateDatabase):
     name = "bigquery"
     compiler = sc.bigquery.compiler
     supports_python_udfs = False
@@ -520,6 +520,10 @@ def disconnect(self) -> None:
 
     def _parse_project_and_dataset(self, dataset) -> tuple[str, str]:
         if isinstance(dataset, sge.Table):
+            if (sg_cat := dataset.args["catalog"]) is not None:
+                sg_cat.args["quoted"] = False
+            if (sg_db := dataset.args["db"]) is not None:
+                sg_db.args["quoted"] = False
             dataset = dataset.sql(self.dialect)
         if not dataset and not self.dataset:
             raise ValueError("Unable to determine BigQuery dataset.")
@@ -582,9 +586,11 @@ def drop_database(
         self.raw_sql(stmt.sql(self.name))
 
     def table(
-        self, name: str, database: str | None = None, schema: str | None = None
+        self,
+        name: str,
+        database: str | None = None,
     ) -> ir.Table:
-        table_loc = self._warn_and_create_table_loc(database, schema)
+        table_loc = self._to_sqlglot_table(database)
         table = sg.parse_one(f"`{name}`", into=sge.Table, read=self.name)
 
         # Bigquery, unlike other backends, had existing support for specifying
@@ -612,10 +618,7 @@ def table(
             else:
                 db = table.db
 
-        database = (
-            sg.table(None, db=db, catalog=catalog, quoted=False).sql(dialect=self.name)
-            or None
-        )
+        database = sg.table(None, db=db, catalog=catalog, quoted=False) or None
 
         project, dataset = self._parse_project_and_dataset(database)
 
@@ -722,7 +725,6 @@ def insert(
         self,
         table_name: str,
         obj: pd.DataFrame | ir.Table | list | dict,
-        schema: str | None = None,
         database: str | None = None,
         overwrite: bool = False,
     ):
@@ -734,15 +736,13 @@ def insert(
             The name of the table to which data needs will be inserted
         obj
             The source data or expression to insert
-        schema
-            The name of the schema that the table is located in
         database
             Name of the attached database that the table is located in.
         overwrite
             If `True` then replace existing contents of table
 
         """
-        table_loc = self._warn_and_create_table_loc(database, schema)
+        table_loc = self._to_sqlglot_table(database)
         catalog, db = self._to_catalog_db_tuple(table_loc)
         if catalog is None:
             catalog = self.current_catalog
@@ -896,7 +896,6 @@ def list_tables(
         self,
         like: str | None = None,
         database: tuple[str, str] | str | None = None,
-        schema: str | None = None,
     ) -> list[str]:
         """List the tables in the database.
 
@@ -924,10 +923,8 @@ def list_tables(
             To specify a table in a separate BigQuery dataset, you can pass in the
             dataset and project as a string `"dataset.project"`, or as a tuple of
             strings `(dataset, project)`.
-        schema
-            [deprecated] The schema (dataset) inside `database` to perform the list against.
         """
-        table_loc = self._warn_and_create_table_loc(database, schema)
+        table_loc = self._to_sqlglot_table(database)
 
         project, dataset = self._parse_project_and_dataset(table_loc)
         dataset_ref = bq.DatasetReference(project, dataset)
@@ -1090,11 +1087,10 @@ def drop_table(
         self,
         name: str,
         *,
-        schema: str | None = None,
         database: tuple[str | str] | str | None = None,
         force: bool = False,
     ) -> None:
-        table_loc = self._warn_and_create_table_loc(database, schema)
+        table_loc = self._to_sqlglot_table(database)
         catalog, db = self._to_catalog_db_tuple(table_loc)
         stmt = sge.Drop(
             kind="TABLE",
@@ -1112,11 +1108,10 @@ def create_view(
         name: str,
         obj: ir.Table,
         *,
-        schema: str | None = None,
         database: str | None = None,
         overwrite: bool = False,
     ) -> ir.Table:
-        table_loc = self._warn_and_create_table_loc(database, schema)
+        table_loc = self._to_sqlglot_table(database)
         catalog, db = self._to_catalog_db_tuple(table_loc)
 
         stmt = sge.Create(
@@ -1137,11 +1132,10 @@ def drop_view(
         self,
         name: str,
         *,
-        schema: str | None = None,
         database: str | None = None,
         force: bool = False,
     ) -> None:
-        table_loc = self._warn_and_create_table_loc(database, schema)
+        table_loc = self._to_sqlglot_table(database)
         catalog, db = self._to_catalog_db_tuple(table_loc)
 
         stmt = sge.Drop(
@@ -1169,32 +1163,6 @@ def _register_udfs(self, expr: ir.Expr) -> None:
     def _safe_raw_sql(self, *args, **kwargs):
         yield self.raw_sql(*args, **kwargs)
 
-    # TODO: remove when the schema kwarg is removed
-    def _warn_and_create_table_loc(self, database=None, schema=None):
-        if schema is not None:
-            self._warn_schema()
-        if database is not None and schema is not None:
-            if isinstance(database, str):
-                table_loc = f"{database}.{schema}"
-            elif isinstance(database, tuple):
-                table_loc = database + schema
-        elif schema is not None:
-            table_loc = schema
-        elif database is not None:
-            table_loc = database
-        else:
-            table_loc = None
-
-        table_loc = self._to_sqlglot_table(table_loc)
-
-        if table_loc is not None:
-            if (sg_cat := table_loc.args["catalog"]) is not None:
-                sg_cat.args["quoted"] = False
-            if (sg_db := table_loc.args["db"]) is not None:
-                sg_db.args["quoted"] = False
-
-        return table_loc
-
 
 def compile(expr, params=None, **kwargs):
     """Compile an expression for BigQuery."""

ibis/backends/bigquery/tests/system/test_client.py

@@ -34,6 +34,16 @@ def test_list_tables(con):
     tables = con.list_tables(like="functional_alltypes")
     assert set(tables) == {"functional_alltypes", "functional_alltypes_parted"}
 
+    pypi_tables = [
+        "external",
+        "native",
+    ]
+
+    assert con.list_tables()
+
+    assert con.list_tables(database="ibis-gbq.pypi") == pypi_tables
+    assert con.list_tables(database=("ibis-gbq", "pypi")) == pypi_tables
+
 
 def test_current_catalog(con):
     assert con.current_catalog == con.billing_project
@@ -386,22 +396,6 @@ def test_create_table_with_options(con):
         con.drop_table(name)
 
 
-def test_list_tables_schema_warning_refactor(con):
-    pypi_tables = [
-        "external",
-        "native",
-    ]
-
-    assert con.list_tables()
-
-    # Warn but succeed for schema list
-    with pytest.raises(FutureWarning):
-        assert con.list_tables(schema="pypi") == pypi_tables
-
-    assert con.list_tables(database="ibis-gbq.pypi") == pypi_tables
-    assert con.list_tables(database=("ibis-gbq", "pypi")) == pypi_tables
-
-
 def test_create_temp_table_from_scratch(project_id, dataset_id):
     con = ibis.bigquery.connect(project_id=project_id, dataset_id=dataset_id)
     name = gen_name("bigquery_temp_table")

ibis/backends/conftest.py

@@ -17,7 +17,6 @@
 from ibis.backends import (
     CanCreateCatalog,
     CanCreateDatabase,
-    CanListSchema,
     _get_backend_names,
 )
 from ibis.conftest import WINDOWS
@@ -424,14 +423,6 @@ def con_no_data(backend_no_data):
     return backend_no_data.connection
 
 
-@pytest.fixture(scope="session")
-def con_list_schema(con):
-    if isinstance(con, CanListSchema):
-        return con
-    else:
-        pytest.skip(f"{con.name} backend cannot create schemas")
-
-
 @pytest.fixture(scope="session")
 def con_create_catalog(con):
     if isinstance(con, CanCreateCatalog):

ibis/backends/datafusion/__init__.py

@@ -22,7 +22,7 @@
 import ibis.expr.schema as sch
 import ibis.expr.types as ir
 from ibis import util
-from ibis.backends import CanCreateCatalog, CanCreateDatabase, CanCreateSchema, NoUrl
+from ibis.backends import CanCreateCatalog, CanCreateDatabase, NoUrl
 from ibis.backends.sql import SQLBackend
 from ibis.backends.sql.compilers.base import C
 from ibis.common.dispatch import lazy_singledispatch
@@ -69,7 +69,7 @@ def as_nullable(dtype: dt.DataType) -> dt.DataType:
         return dtype.copy(nullable=True)
 
 
-class Backend(SQLBackend, CanCreateCatalog, CanCreateDatabase, CanCreateSchema, NoUrl):
+class Backend(SQLBackend, CanCreateCatalog, CanCreateDatabase, NoUrl):
     name = "datafusion"
     supports_arrays = True
     compiler = sc.datafusion.compiler
@@ -674,8 +674,10 @@ def create_table(
         return self.table(name, database=database)
 
     def truncate_table(
-        self, name: str, database: str | None = None, schema: str | None = None
-    ) -> None:
+        self,
+        name: str,
+        database: str | None = None,
+    ):
         """Delete all rows from a table.
 
         Parameters
@@ -684,14 +686,12 @@ def truncate_table(
             Table name
         database
             Database name
-        schema
-            Schema name
 
         """
         # datafusion doesn't support `TRUNCATE TABLE` so we use `DELETE FROM`
         #
         # however datafusion as of 34.0.0 doesn't implement DELETE DML yet
-        table_loc = self._warn_and_create_table_loc(database, schema)
+        table_loc = self._to_sqlglot_table(database)
         catalog, db = self._to_catalog_db_tuple(table_loc)
 
         ident = sg.table(name, db=db, catalog=catalog).sql(self.dialect)