Remove unused code: schema fetching via dialects

The schemas are currently fetching via `Database.select_table_schema()` & `Database.query_table_schema()`, so this dialectic code is unused.

It is also questionable that the schema fetching must be done via dialects at all: some databases can provide an API (e.g. HTTP API) to fetch the database metadata, such as Databricks. So it belongs to the database, not to dialects.

Author: Sergey Vasilyev

data_diff/abcs/mixins.py

@@ -122,25 +122,6 @@ def md5_as_int(self, s: str) -> str:
         "Provide SQL for computing md5 and returning an int"
 
 
-@attrs.define(frozen=False)
-class AbstractMixin_Schema(AbstractMixin):
-    """Methods for querying the database schema
-
-    TODO: Move AbstractDatabase.query_table_schema() and friends over here
-    """
-
-    def table_information(self) -> Compilable:
-        "Query to return a table of schema information about existing tables"
-        raise NotImplementedError()
-
-    @abstractmethod
-    def list_tables(self, table_schema: str, like: Compilable = None) -> Compilable:
-        """Query to select the list of tables in the schema. (query return type: table[str])
-
-        If 'like' is specified, the value is applied to the table name, using the 'like' operator.
-        """
-
-
 @attrs.define(frozen=False)
 class AbstractMixin_OptimizerHints(AbstractMixin):
     @abstractmethod

data_diff/databases/base.py

@@ -75,7 +75,6 @@
 )
 from data_diff.abcs.mixins import Compilable
 from data_diff.abcs.mixins import (
-    AbstractMixin_Schema,
     AbstractMixin_NormalizeValue,
     AbstractMixin_OptimizerHints,
 )
@@ -199,23 +198,6 @@ def apply_query(callback: Callable[[str], Any], sql_code: Union[str, ThreadLocal
         return callback(sql_code)
 
 
-@attrs.define(frozen=False)
-class Mixin_Schema(AbstractMixin_Schema):
-    def table_information(self) -> Compilable:
-        return table("information_schema", "tables")
-
-    def list_tables(self, table_schema: str, like: Compilable = None) -> Compilable:
-        return (
-            self.table_information()
-            .where(
-                this.table_schema == table_schema,
-                this.table_name.like(like) if like is not None else SKIP,
-                this.table_type == "BASE TABLE",
-            )
-            .select(this.table_name)
-        )
-
-
 @attrs.define(frozen=False)
 class Mixin_OptimizerHints(AbstractMixin_OptimizerHints):
     def optimizer_hints(self, hints: str) -> str:
@@ -1019,10 +1001,6 @@ def _refine_coltypes(self, table_path: DbPath, col_dict: Dict[str, ColType], whe
                         assert col_name in col_dict
                         col_dict[col_name] = String_VaryingAlphanum()
 
-    # @lru_cache()
-    # def get_table_schema(self, path: DbPath) -> Dict[str, ColType]:
-    #     return self.query_table_schema(path)
-
     def _normalize_table_path(self, path: DbPath) -> DbPath:
         if len(path) == 1:
             return self.default_schema, path[0]
@@ -1056,9 +1034,6 @@ def close(self):
         self.is_closed = True
         return super().close()
 
-    def list_tables(self, tables_like, schema=None):
-        return self.query(self.dialect.list_tables(schema or self.default_schema, tables_like))
-
     @property
     @abstractmethod
     def dialect(self) -> BaseDialect:

data_diff/databases/bigquery.py

@@ -23,7 +23,6 @@
 from data_diff.abcs.mixins import (
     AbstractMixin_MD5,
     AbstractMixin_NormalizeValue,
-    AbstractMixin_Schema,
 )
 from data_diff.abcs.compiler import Compilable
 from data_diff.queries.api import this, table, SKIP, code
@@ -62,7 +61,7 @@ def import_bigquery_service_account_impersonation():
 
 
 @attrs.define(frozen=False)
-class Dialect(BaseDialect, AbstractMixin_Schema, AbstractMixin_MD5, AbstractMixin_NormalizeValue):
+class Dialect(BaseDialect, AbstractMixin_MD5, AbstractMixin_NormalizeValue):
     name = "BigQuery"
     ROUNDS_ON_PREC_LOSS = False  # Technically BigQuery doesn't allow implicit rounding or truncation
     TYPE_CLASSES = {
@@ -183,17 +182,6 @@ def normalize_struct(self, value: str, _coltype: Struct) -> str:
         # match on both sides: i.e. have properly ordered keys, same spacing, same quotes, etc.
         return f"to_json_string({value})"
 
-    def list_tables(self, table_schema: str, like: Compilable = None) -> Compilable:
-        return (
-            table(table_schema, "INFORMATION_SCHEMA", "TABLES")
-            .where(
-                this.table_schema == table_schema,
-                this.table_name.like(like) if like is not None else SKIP,
-                this.table_type == "BASE TABLE",
-            )
-            .select(this.table_name)
-        )
-
 
 @attrs.define(frozen=False, init=False, kw_only=True)
 class BigQuery(Database):

data_diff/databases/duckdb.py

@@ -30,9 +30,7 @@
     TIMESTAMP_PRECISION_POS,
     CHECKSUM_OFFSET,
 )
-from data_diff.databases.base import MD5_HEXDIGITS, CHECKSUM_HEXDIGITS, Mixin_Schema
-from data_diff.queries.ast_classes import ITable
-from data_diff.queries.api import code
+from data_diff.databases.base import MD5_HEXDIGITS, CHECKSUM_HEXDIGITS
 
 
 @import_helper("duckdb")
@@ -43,7 +41,7 @@ def import_duckdb():
 
 
 @attrs.define(frozen=False)
-class Dialect(BaseDialect, Mixin_Schema, AbstractMixin_MD5, AbstractMixin_NormalizeValue):
+class Dialect(BaseDialect, AbstractMixin_MD5, AbstractMixin_NormalizeValue):
     name = "DuckDB"
     ROUNDS_ON_PREC_LOSS = False
     SUPPORTS_PRIMARY_KEY = True

data_diff/databases/mssql.py

@@ -13,7 +13,6 @@
     ConnectError,
     BaseDialect,
 )
-from data_diff.databases.base import Mixin_Schema
 from data_diff.abcs.database_types import (
     JSON,
     NumericType,
@@ -40,7 +39,6 @@ def import_mssql():
 @attrs.define(frozen=False)
 class Dialect(
     BaseDialect,
-    Mixin_Schema,
     Mixin_OptimizerHints,
     AbstractMixin_MD5,
     AbstractMixin_NormalizeValue,

data_diff/databases/mysql.py

@@ -31,7 +31,6 @@
     CHECKSUM_HEXDIGITS,
     TIMESTAMP_PRECISION_POS,
     CHECKSUM_OFFSET,
-    Mixin_Schema,
 )
 
 
@@ -45,7 +44,6 @@ def import_mysql():
 @attrs.define(frozen=False)
 class Dialect(
     BaseDialect,
-    Mixin_Schema,
     Mixin_OptimizerHints,
     AbstractMixin_MD5,
     AbstractMixin_NormalizeValue,

data_diff/databases/oracle.py

@@ -16,9 +16,7 @@
     TimestampTZ,
     FractionalType,
 )
-from data_diff.abcs.mixins import AbstractMixin_MD5, AbstractMixin_NormalizeValue, AbstractMixin_Schema
-from data_diff.abcs.compiler import Compilable
-from data_diff.queries.api import this, table, SKIP
+from data_diff.abcs.mixins import AbstractMixin_MD5, AbstractMixin_NormalizeValue
 from data_diff.databases.base import (
     BaseDialect,
     Mixin_OptimizerHints,
@@ -46,7 +44,6 @@ def import_oracle():
 class Dialect(
     BaseDialect,
     Mixin_OptimizerHints,
-    AbstractMixin_Schema,
     AbstractMixin_MD5,
     AbstractMixin_NormalizeValue,
 ):
@@ -162,16 +159,6 @@ def normalize_number(self, value: str, coltype: FractionalType) -> str:
             format_str += "0." + "9" * (coltype.precision - 1) + "0"
         return f"to_char({value}, '{format_str}')"
 
-    def list_tables(self, table_schema: str, like: Compilable = None) -> Compilable:
-        return (
-            table("ALL_TABLES")
-            .where(
-                this.OWNER == table_schema,
-                this.TABLE_NAME.like(like) if like is not None else SKIP,
-            )
-            .select(table_name=this.TABLE_NAME)
-        )
-
 
 @attrs.define(frozen=False, init=False, kw_only=True)
 class Oracle(ThreadedDatabase):

data_diff/databases/postgresql.py

@@ -19,7 +19,7 @@
     Date,
 )
 from data_diff.abcs.mixins import AbstractMixin_MD5, AbstractMixin_NormalizeValue
-from data_diff.databases.base import BaseDialect, ThreadedDatabase, import_helper, ConnectError, Mixin_Schema
+from data_diff.databases.base import BaseDialect, ThreadedDatabase, import_helper, ConnectError
 from data_diff.databases.base import (
     MD5_HEXDIGITS,
     CHECKSUM_HEXDIGITS,
@@ -40,7 +40,7 @@ def import_postgresql():
 
 
 @attrs.define(frozen=False)
-class PostgresqlDialect(BaseDialect, Mixin_Schema, AbstractMixin_MD5, AbstractMixin_NormalizeValue):
+class PostgresqlDialect(BaseDialect, AbstractMixin_MD5, AbstractMixin_NormalizeValue):
     name = "PostgreSQL"
     ROUNDS_ON_PREC_LOSS = True
     SUPPORTS_PRIMARY_KEY = True

data_diff/databases/presto.py

@@ -27,7 +27,6 @@
     Database,
     import_helper,
     ThreadLocalInterpreter,
-    Mixin_Schema,
 )
 from data_diff.databases.base import (
     MD5_HEXDIGITS,
@@ -53,7 +52,7 @@ def import_presto():
     return prestodb
 
 
-class Dialect(BaseDialect, Mixin_Schema, AbstractMixin_MD5, AbstractMixin_NormalizeValue):
+class Dialect(BaseDialect, AbstractMixin_MD5, AbstractMixin_NormalizeValue):
     name = "Presto"
     ROUNDS_ON_PREC_LOSS = True
     TYPE_CLASSES = {

data_diff/databases/snowflake.py

@@ -18,7 +18,6 @@
 from data_diff.abcs.mixins import (
     AbstractMixin_MD5,
     AbstractMixin_NormalizeValue,
-    AbstractMixin_Schema,
 )
 from data_diff.abcs.compiler import Compilable
 from data_diff.queries.api import table, this, SKIP, code
@@ -42,7 +41,7 @@ def import_snowflake():
     return snowflake, serialization, default_backend
 
 
-class Dialect(BaseDialect, AbstractMixin_Schema, AbstractMixin_MD5, AbstractMixin_NormalizeValue):
+class Dialect(BaseDialect, AbstractMixin_MD5, AbstractMixin_NormalizeValue):
     name = "Snowflake"
     ROUNDS_ON_PREC_LOSS = False
     TYPE_CLASSES = {
@@ -69,9 +68,6 @@ def quote(self, s: str):
     def to_string(self, s: str):
         return f"cast({s} as string)"
 
-    def table_information(self) -> Compilable:
-        return table("INFORMATION_SCHEMA", "TABLES")
-
     def set_timezone_to_utc(self) -> str:
         return "ALTER SESSION SET TIMEZONE = 'UTC'"
 
@@ -100,20 +96,6 @@ def normalize_number(self, value: str, coltype: FractionalType) -> str:
     def normalize_boolean(self, value: str, _coltype: Boolean) -> str:
         return self.to_string(f"{value}::int")
 
-    def table_information(self) -> Compilable:
-        return table("INFORMATION_SCHEMA", "TABLES")
-
-    def list_tables(self, table_schema: str, like: Compilable = None) -> Compilable:
-        return (
-            self.table_information()
-            .where(
-                this.TABLE_SCHEMA == table_schema,
-                this.TABLE_NAME.like(like) if like is not None else SKIP,
-                this.TABLE_TYPE == "BASE TABLE",
-            )
-            .select(table_name=this.TABLE_NAME)
-        )
-
 
 @attrs.define(frozen=False, init=False, kw_only=True)
 class Snowflake(Database):

data_diff/databases/vertica.py

@@ -27,9 +27,7 @@
     Boolean,
     ColType_UUID,
 )
-from data_diff.abcs.mixins import AbstractMixin_MD5, AbstractMixin_NormalizeValue, AbstractMixin_Schema
-from data_diff.abcs.compiler import Compilable
-from data_diff.queries.api import table, this, SKIP
+from data_diff.abcs.mixins import AbstractMixin_MD5, AbstractMixin_NormalizeValue
 
 
 @import_helper("vertica")
@@ -39,7 +37,7 @@ def import_vertica():
     return vertica_python
 
 
-class Dialect(BaseDialect, AbstractMixin_MD5, AbstractMixin_NormalizeValue, AbstractMixin_Schema):
+class Dialect(BaseDialect, AbstractMixin_MD5, AbstractMixin_NormalizeValue):
     name = "Vertica"
     ROUNDS_ON_PREC_LOSS = True
 
@@ -131,19 +129,6 @@ def normalize_uuid(self, value: str, _coltype: ColType_UUID) -> str:
     def normalize_boolean(self, value: str, _coltype: Boolean) -> str:
         return self.to_string(f"cast ({value} as int)")
 
-    def table_information(self) -> Compilable:
-        return table("v_catalog", "tables")
-
-    def list_tables(self, table_schema: str, like: Compilable = None) -> Compilable:
-        return (
-            self.table_information()
-            .where(
-                this.table_schema == table_schema,
-                this.table_name.like(like) if like is not None else SKIP,
-            )
-            .select(this.table_name)
-        )
-
 
 @attrs.define(frozen=False, init=False, kw_only=True)
 class Vertica(ThreadedDatabase):

tests/test_database.py

@@ -63,44 +63,6 @@ def test_bad_uris(self):
         self.assertRaises(ValueError, connect, "snowflake://user:pass@foo/bar/TEST1?warehouse=ha&schema=dup")
 
 
-@test_each_database
-class TestSchema(unittest.TestCase):
-    def test_table_list(self):
-        name = "tbl_" + random_table_suffix()
-        db = get_conn(self.db_cls)
-        tbl = table(db.dialect.parse_table_name(name), schema={"id": int})
-        q = db.dialect.list_tables(db.default_schema, name)
-        assert not db.query(q)
-
-        db.query(tbl.create())
-        self.assertEqual(db.query(q, List[str]), [name])
-
-        db.query(tbl.drop())
-        assert not db.query(q)
-
-    def test_type_mapping(self):
-        name = "tbl_" + random_table_suffix()
-        db = get_conn(self.db_cls)
-        tbl = table(
-            db.dialect.parse_table_name(name),
-            schema={
-                "int": int,
-                "float": float,
-                "datetime": datetime,
-                "str": str,
-                "bool": bool,
-            },
-        )
-        q = db.dialect.list_tables(db.default_schema, name)
-        assert not db.query(q)
-
-        db.query(tbl.create())
-        self.assertEqual(db.query(q, List[str]), [name])
-
-        db.query(tbl.drop())
-        assert not db.query(q)
-
-
 @test_each_database
 class TestQueries(unittest.TestCase):
     def test_current_timestamp(self):