maintenance: bump version to 1.7.0

maintenance: update changelog with release notes
maintenance: update __version__ in dataset/__init__

.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 1.5.2
+current_version = 1.6.2
 tag_name = {new_version}
 commit = True
 tag = True

CHANGELOG.md

@@ -1,5 +1,13 @@
 # dataset ChangeLog
 
+
+## Version 1.7.0 - 2025-11-09
+
+Improved database interactions for faster and more reliable data operations. Updated dependencies and tests to support these changes.
+
+**Changes:** 5 files, 68 lines
+**Languages:** Python (68 lines)
+
 *The changelog has only been started with version 0.3.12, previous
 changes must be reconstructed from revision history.*
 

dataset/__init__.py

@@ -11,7 +11,7 @@ warnings.filterwarnings(
 )
 
 __all__ = ["Database", "Table", "connect"]
-__version__ = "1.5.2"
+__version__ = "1.7.0"
 
 
 def connect(

dataset/database.py

@@ -106,14 +106,12 @@ class Database(object):
     @property
     def metadata(self):
         """Return a SQLAlchemy schema cache object."""
-        return MetaData(schema=self.schema, bind=self.executable)
+        return MetaData(schema=self.schema)
 
     @property
     def in_transaction(self):
         """Check if this database is in a transactional context."""
-        if not hasattr(self.local, "tx"):
+        return self.executable.in_transaction()
-            return False
-        return len(self.local.tx) > 0
 
     def _flush_tables(self):
         """Clear the table metadata after transaction rollbacks."""
@@ -127,7 +125,10 @@ class Database(object):
         """
         if not hasattr(self.local, "tx"):
             self.local.tx = []
-        self.local.tx.append(self.executable.begin())
+        if self.in_transaction:
+            self.local.tx.append(self.executable.begin_nested())
+        else:
+            self.local.tx.append(self.executable.begin())
 
     def commit(self):
         """Commit the current transaction.

dataset/table.py

@@ -116,7 +116,7 @@ class Table(object):
         Returns the inserted row's primary key.
         """
         row = self._sync_columns(row, ensure, types=types)
-        res = self.db.executable.execute(self.table.insert(row))
+        res = self.db.executable.execute(self.table.insert().values(row))
         if len(res.inserted_primary_key) > 0:
             return res.inserted_primary_key[0]
         return True
@@ -181,7 +181,7 @@ class Table(object):
             # Insert when chunk_size is fulfilled or this is the last row
             if len(chunk) == chunk_size or index == len(rows) - 1:
                 chunk = pad_chunk_columns(chunk, columns)
-                self.table.insert().execute(chunk)
+                self.db.executable.execute(self.table.insert(), chunk)
                 chunk = []
 
     def update(self, row, keys, ensure=None, types=None, return_count=False):
@@ -206,7 +206,7 @@ class Table(object):
         clause = self._args_to_clause(args)
         if not len(row):
             return self.count(clause)
-        stmt = self.table.update(whereclause=clause, values=row)
+        stmt = self.table.update().where(clause).values(row)
         rp = self.db.executable.execute(stmt)
         if rp.supports_sane_rowcount():
             return rp.rowcount
@@ -241,9 +241,10 @@ class Table(object):
             # Update when chunk_size is fulfilled or this is the last row
             if len(chunk) == chunk_size or index == len(rows) - 1:
                 cl = [self.table.c[k] == bindparam("_%s" % k) for k in keys]
-                stmt = self.table.update(
+                stmt = (
-                    whereclause=and_(True, *cl),
+                    self.table.update()
-                    values={col: bindparam(col, required=False) for col in columns},
+                    .where(and_(True, *cl))
+                    .values({col: bindparam(col, required=False) for col in columns})
                 )
                 self.db.executable.execute(stmt, chunk)
                 chunk = []
@@ -293,7 +294,7 @@ class Table(object):
         if not self.exists:
             return False
         clause = self._args_to_clause(filters, clauses=clauses)
-        stmt = self.table.delete(whereclause=clause)
+        stmt = self.table.delete().where(clause)
         rp = self.db.executable.execute(stmt)
         return rp.rowcount > 0
 
@@ -303,7 +304,10 @@ class Table(object):
             self._columns = None
             try:
                 self._table = SQLATable(
-                    self.name, self.db.metadata, schema=self.db.schema, autoload=True
+                    self.name,
+                    self.db.metadata,
+                    schema=self.db.schema,
+                    autoload_with=self.db.executable,
                 )
             except NoSuchTableError:
                 self._table = None
@@ -353,7 +357,7 @@ class Table(object):
                 self._threading_warn()
                 for column in columns:
                     if not self.has_column(column.name):
-                        self.db.op.add_column(self.name, column, self.db.schema)
+                        self.db.op.add_column(self.name, column, schema=self.db.schema)
                 self._reflect_table()
 
     def _sync_columns(self, row, ensure, types=None):
@@ -509,7 +513,7 @@ class Table(object):
                 return
 
             self._threading_warn()
-            self.db.op.drop_column(self.table.name, name, self.table.schema)
+            self.db.op.drop_column(self.table.name, name, schema=self.table.schema)
             self._reflect_table()
 
     def drop(self):
@@ -625,7 +629,7 @@ class Table(object):
 
         order_by = self._args_to_order_by(order_by)
         args = self._args_to_clause(kwargs, clauses=_clauses)
-        query = self.table.select(whereclause=args, limit=_limit, offset=_offset)
+        query = self.table.select().where(args).limit(_limit).offset(_offset)
         if len(order_by):
             query = query.order_by(*order_by)
 
@@ -666,7 +670,7 @@ class Table(object):
             return 0
 
         args = self._args_to_clause(kwargs, clauses=_clauses)
-        query = select([func.count()], whereclause=args)
+        query = select(func.count()).where(args)
         query = query.select_from(self.table)
         rp = self.db.executable.execute(query)
         return rp.fetchone()[0]
@@ -675,7 +679,7 @@ class Table(object):
         """Return the number of rows in the table."""
         return self.count()
 
-    def distinct(self, *args, **_filter):
+    def distinct(self, *args, **kwargs):
         """Return all the unique (distinct) values for the given ``columns``.
         ::
 
@@ -699,15 +703,19 @@ class Table(object):
                     raise DatasetException("No such column: %s" % column)
                 columns.append(self.table.c[column])
 
-        clause = self._args_to_clause(_filter, clauses=clauses)
+        _limit = kwargs.pop("_limit", None)
+        _offset = kwargs.pop("_offset", 0)
+        clause = self._args_to_clause(kwargs, clauses=clauses)
         if not len(columns):
             return iter([])
 
-        q = expression.select(
+        q = (
-            columns,
+            expression.select(*columns)
-            distinct=True,
+            .distinct()
-            whereclause=clause,
+            .where(clause)
-            order_by=[c.asc() for c in columns],
+            .limit(_limit)
+            .offset(_offset)
+            .order_by(*[c.asc() for c in columns])
         )
         return self.db.query(q)
 

dataset/util.py

@@ -15,7 +15,6 @@ try:
             return None
         return row_type(row._mapping.items())
 
-
 except ImportError:
     # SQLAlchemy < 1.4.0, no _mapping.
 

docs/conf.py

@@ -45,9 +45,9 @@ copyright = u"2013-2021, Friedrich Lindenberg, Gregor Aisch, Stefan Wehrmeyer"
 # built documents.
 #
 # The short X.Y version.
-version = "1.5.2"
+version = "1.6.2"
 # The full version, including alpha/beta/rc tags.
-release = "1.5.2"
+release = "1.6.2"
 
 # There are two options for replacing |today|: either, you set today to some
 # non-false value, then it is used:

setup.py

@@ -5,7 +5,7 @@ with open("README.md") as f:
 
 setup(
     name="dataset",
-    version="1.5.2",
+    version="1.7.0",
     description="Toolkit for Python-based database access.",
     long_description=long_description,
     long_description_content_type="text/markdown",
@@ -30,7 +30,7 @@ setup(
     include_package_data=False,
     zip_safe=False,
     install_requires=[
-        "sqlalchemy >= 1.3.2, < 2.0.0",
+        "sqlalchemy >= 2.0.0",
         "alembic >= 0.6.2",
         "banal >= 1.0.1",
     ],

test/test_dataset.py

@@ -45,7 +45,7 @@ class DatabaseTestCase(unittest.TestCase):
         if "sqlite" in self.db.engine.dialect.dbapi.__name__:
             return
         table = self.db.create_table("foo_no_id", primary_id=False)
-        assert table.table.exists()
+        assert self.db.has_table(table.table.name)
         assert len(table.table.columns) == 0, table.table.columns
 
     def test_create_table_custom_id1(self):
@@ -83,7 +83,7 @@ class DatabaseTestCase(unittest.TestCase):
     def test_create_table_shorthand1(self):
         pid = "int_id"
         table = self.db.get_table("foo5", pid)
-        assert table.table.exists
+        assert self.db.has_table(table.table.name)
         assert len(table.table.columns) == 1, table.table.columns
         assert pid in table.table.c, table.table.c
 
@@ -98,7 +98,7 @@ class DatabaseTestCase(unittest.TestCase):
         table = self.db.get_table(
             "foo6", primary_id=pid, primary_type=self.db.types.string(255)
         )
-        assert table.table.exists
+        assert self.db.has_table(table.table.name)
         assert len(table.table.columns) == 1, table.table.columns
         assert pid in table.table.c, table.table.c
 
@@ -384,6 +384,16 @@ class TableTestCase(unittest.TestCase):
         assert len(x) == 3, x
         x = list(self.tbl.distinct("temperature", place=["B€rkeley", "G€lway"]))
         assert len(x) == 6, x
+        x = list(
+            self.tbl.distinct("temperature", _limit=3, place=["B€rkeley", "G€lway"])
+        )
+        assert len(x) == 3, x
+        x = list(
+            self.tbl.distinct(
+                "temperature", _limit=6, _offset=1, place=["B€rkeley", "G€lway"]
+            )
+        )
+        assert len(x) == 5, x
 
     def test_insert_many(self):
         data = TEST_DATA * 100