diff --git a/docs/source/changelog.md b/docs/source/changelog.md
index dbc26dbc..e8b99a8a 100644
--- a/docs/source/changelog.md
+++ b/docs/source/changelog.md
@@ -2,6 +2,7 @@
 
 ## 0.6.4 (2025-12-18)
 - fix col.map() Null-type handling and string operations
+- fix print display of short tables
 
 ## 0.6.3 (2025-12-17)
 - implement `tbl1 >> union(tbl2)` and `union(tbl1,tbl2)`
diff --git a/pixi.toml b/pixi.toml
index 911191bd..798e4e7d 100644
--- a/pixi.toml
+++ b/pixi.toml
@@ -26,6 +26,7 @@ pixi-pycharm = ">=0.0.6"
 pytest = ">=7.1.2"
 pytest-xdist = ">=2.5.0"
 structlog = ">=25.4.0,<26"
+filelock = ">=3.20.0"
 
 [feature.release.dependencies]
 hatch = ">=1.12.0"
diff --git a/src/pydiverse/transform/_internal/pipe/table.py b/src/pydiverse/transform/_internal/pipe/table.py
index d061262d..3ed8f3fd 100644
--- a/src/pydiverse/transform/_internal/pipe/table.py
+++ b/src/pydiverse/transform/_internal/pipe/table.py
@@ -275,7 +275,7 @@ def __repr__(self) -> str:
                 query = "\n\nQuery:\n" + query
         except Exception:
             query = ""
-        return res + str(df).split("\n", 1)[1]
+        return res + str(df).split("\n", 1)[1] + query
 
     def __dir__(self) -> list[str]:
         return [name for name in self._cache.name_to_uuid.keys()]
@@ -314,7 +314,9 @@ def get_head_tail(tbl: Table) -> tuple[pl.DataFrame, int]:
 
     height = tbl >> summarize(num_rows=pdt.count()) >> export(pdt.Scalar)
     tbl_rows = int(pl.Config.state().get("POLARS_FMT_MAX_ROWS") or 10)
-    head_tail_len = tbl_rows // 2 + 1
+    if height <= tbl_rows:
+        return tbl >> export(pdt.Polars), height
+    head_tail_len = tbl_rows // 2
 
     # Only export the first and last few rows.
     head: pl.DataFrame = tbl >> slice_head(head_tail_len) >> export(pdt.Polars)
diff --git a/src/pydiverse/transform/_internal/pipe/verbs.py b/src/pydiverse/transform/_internal/pipe/verbs.py
index 6b03c8b9..8f755009 100644
--- a/src/pydiverse/transform/_internal/pipe/verbs.py
+++ b/src/pydiverse/transform/_internal/pipe/verbs.py
@@ -1360,42 +1360,6 @@ def _union_impl(
     *,
     distinct: bool = False,
 ) -> Table:
-    """
-    Unions two tables by stacking rows vertically.
-
-    The left table in the union comes through the pipe `>>` operator from the
-    left.
-
-    :param right:
-        The right table to union with.
-
-    :param distinct:
-        If ``True``, performs UNION (removes duplicates). If ``False``,
-        performs UNION ALL (keeps duplicates).
-
-    Note
-    ----
-    Both tables must have the same number of columns with compatible types.
-    Column names must match between the two tables.
-
-    Examples
-    --------
-    >>> t1 = pdt.Table({"a": [1, 2, 3], "b": [4, 5, 6]}, name="t1")
-    >>> t2 = pdt.Table({"a": [7, 8], "b": [9, 10]}, name="t2")
-    >>> t1 >> union(t2) >> show()
-    shape: (5, 2)
-    ┌─────┬─────┐
-    │ a   ┆ b   │
-    │ --- ┆ --- │
-    │ i64 ┆ i64 │
-    ╞═════╪═════╡
-    │ 1   ┆ 4   │
-    │ 2   ┆ 5   │
-    │ 3   ┆ 6   │
-    │ 7   ┆ 9   │
-    │ 8   ┆ 10  │
-    └─────┴─────┘
-    """
     errors.check_arg_type(Table, "union", "right", right)
 
     if left._cache.backend != right._cache.backend:
@@ -1457,11 +1421,15 @@ def union(
     """
     Unions two tables by stacking rows vertically.
 
-    The left table in the union comes through the pipe `>>` operator from the
+    The left table in the union may come through the pipe `>>` operator from the
     left.
 
+    :param left_or_right:
+        In case of tbl1 >> union(tbl2) this is the right table (tbl2).
+        In case of union(tbl1, tbl2) this is the left table (tbl1).
+
     :param right:
-        The right table to union with.
+        The right table to union with in case of union(tbl1, tbl2).
 
     :param distinct:
         If ``True``, performs UNION (removes duplicates). If ``False``,
@@ -1490,7 +1458,7 @@ def union(
     │ 8   ┆ 10  │
     └─────┴─────┘
 
-    You can also call it directly:
+    You can also call it in this form:
     >>> union(t1, t2) >> show()
     """
     # If called with two arguments directly (union(tbl1, tbl2)), return Table directly
diff --git a/tests/test_backend_equivalence/conftest.py b/tests/test_backend_equivalence/conftest.py
index ecff0d2e..c74f4c5f 100644
--- a/tests/test_backend_equivalence/conftest.py
+++ b/tests/test_backend_equivalence/conftest.py
@@ -8,6 +8,7 @@
 
 from tests.fixtures.backend import BACKEND_MARKS, flatten
 from tests.util.backend import BACKEND_TABLES
+from tests.util.filelock import lock
 
 dataframes = {
     "df1": pl.DataFrame(
@@ -310,16 +311,23 @@ def pytest_generate_tests(metafunc: pytest.Metafunc):
     return metafunc.parametrize(param_names, params, indirect=True)
 
 
+@pytest.fixture(scope="session")
+def locked():
+    with lock("/tmp/transform/lock"):
+        yield True
+
+
 def generate_df_fixtures():
     def gen_fixture(table_name):
         @pytest.fixture(scope="function")
-        def table_fixture(request):
+        def table_fixture(request, locked):
+            _ = locked
             df = dataframes[table_name]
             name = table_name
 
             table_x = BACKEND_TABLES[request.param[0]](df, name)
             table_y = BACKEND_TABLES[request.param[1]](df, name)
-            return table_x, table_y
+            yield table_x, table_y
 
         return table_fixture
 
diff --git a/tests/test_backend_equivalence/test_print.py b/tests/test_backend_equivalence/test_print.py
index 307d83d0..88cb0352 100644
--- a/tests/test_backend_equivalence/test_print.py
+++ b/tests/test_backend_equivalence/test_print.py
@@ -6,7 +6,7 @@
 
 def test_preview_print(df3, df4, df_strings):
     def data_part(p: str):
-        return p.split("\n", 1)[1]
+        return p.split("\n", 1)[1].split("\n\n", 1)[0]
 
     assert data_part(str(df3[0])) == data_part(str(df3[1]))
     assert data_part(str(df_strings[0])) == data_part(str(df_strings[1]))
diff --git a/tests/test_backend_equivalence/test_union.py b/tests/test_backend_equivalence/test_union.py
index 3f431992..bbfe6ec7 100644
--- a/tests/test_backend_equivalence/test_union.py
+++ b/tests/test_backend_equivalence/test_union.py
@@ -7,6 +7,7 @@
 from pydiverse.transform.extended import *
 from tests.fixtures.backend import skip_backends
 from tests.util import assert_result_equal
+from tests.util.filelock import lock
 
 
 def test_union_basic(df3, df4):
@@ -189,13 +190,15 @@ def test_union_error_different_backends():
 
     # Create tables with different backends
     polars_tbl = pdt.Table(pl.DataFrame({"a": [1, 2]}))
-    engine = sqa.create_engine("sqlite:///:memory:")
-    pl.DataFrame({"a": [1, 2]}).write_database("test", engine, if_table_exists="replace")
-    sql_tbl = pdt.Table("test", pdt.SqlAlchemy(engine))
-
-    # Should raise TypeError
-    with pytest.raises(TypeError, match="cannot union two tables with different backends"):
-        polars_tbl >> union(sql_tbl)
+    file = "/tmp/transform/test2.sqlite"
+    with lock(file):
+        engine = sqa.create_engine("sqlite:///" + file)
+        pl.DataFrame({"a": [1, 2]}).write_database("test", engine, if_table_exists="replace")
+        sql_tbl = pdt.Table("test", pdt.SqlAlchemy(engine))
+
+        # Should raise TypeError
+        with pytest.raises(TypeError, match="cannot union two tables with different backends"):
+            polars_tbl >> union(sql_tbl)
 
 
 def test_union_error_grouped_table(df3, df4):
diff --git a/tests/test_sql_table.py b/tests/test_sql_table.py
index 84542422..1b568dc9 100644
--- a/tests/test_sql_table.py
+++ b/tests/test_sql_table.py
@@ -8,6 +8,7 @@
 import pydiverse.transform as pdt
 from pydiverse.transform.extended import *
 from tests.util import assert_equal
+from tests.util.filelock import lock
 
 df1 = pl.DataFrame(
     {
@@ -60,21 +61,23 @@
 
 @pytest.fixture
 def engine():
-    engine = sqa.create_engine("sqlite:///:memory:")
-    # engine = sqa.create_engine("postgresql://sqa:Pydiverse23@127.0.0.1:6543")
-    # engine = sqa.create_engine(
-    #     "mssql+pyodbc://sqa:PydiQuant27@127.0.0.1:1433"
-    #     "/master?driver=ODBC+Driver+18+for+SQL+Server&encrypt=no"
-    # )
-    # engine = sqa.create_engine("db2+ibm_db://db2inst1:password@localhost:50000/testdb")
-
-    df1.write_database("df1", engine, if_table_exists="replace")
-    df2.write_database("df2", engine, if_table_exists="replace")
-    df3.write_database("df3", engine, if_table_exists="replace")
-    df4.write_database("df4", engine, if_table_exists="replace")
-    df_left.write_database("df_left", engine, if_table_exists="replace")
-    df_right.write_database("df_right", engine, if_table_exists="replace")
-    return engine
+    file = "/tmp/transform/test_pdt.sqlite"
+    with lock(file):
+        engine = sqa.create_engine("sqlite:///" + file)
+        # engine = sqa.create_engine("postgresql://sqa:Pydiverse23@127.0.0.1:6543")
+        # engine = sqa.create_engine(
+        #     "mssql+pyodbc://sqa:PydiQuant27@127.0.0.1:1433"
+        #     "/master?driver=ODBC+Driver+18+for+SQL+Server&encrypt=no"
+        # )
+        # engine = sqa.create_engine("db2+ibm_db://db2inst1:password@localhost:50000/testdb")
+
+        df1.write_database("df1", engine, if_table_exists="replace")
+        df2.write_database("df2", engine, if_table_exists="replace")
+        df3.write_database("df3", engine, if_table_exists="replace")
+        df4.write_database("df4", engine, if_table_exists="replace")
+        df_left.write_database("df_left", engine, if_table_exists="replace")
+        df_right.write_database("df_right", engine, if_table_exists="replace")
+        yield engine
 
 
 @pytest.fixture
diff --git a/tests/util/backend.py b/tests/util/backend.py
index 6fdda9ec..1889ec26 100644
--- a/tests/util/backend.py
+++ b/tests/util/backend.py
@@ -79,12 +79,12 @@ def sql_table(
 
 @_cached_table
 def sqlite_table(df: pl.DataFrame, name: str):
-    return sql_table(df, name, "sqlite:///:memory:")
+    return sql_table(df, name, "sqlite:////tmp/transform/test.sqlite")
 
 
 @_cached_table
 def duckdb_table(df: pl.DataFrame, name: str):
-    return sql_table(df, name, "duckdb:///:memory:")
+    return sql_table(df, name, "duckdb:////tmp/transform/test.duckdb")
 
 
 @_cached_table
@@ -96,7 +96,9 @@ def duckdb_parquet_table(df: pl.DataFrame, name: str):
     if "duckdb_parquet" in _sql_engine_cache:
         engine = _sql_engine_cache["duckdb_parquet"]
     else:
-        engine = sqa.create_engine("duckdb:///:memory:")
+        file = "/tmp/transform/test_parquet.duckdb"
+        Path(file).parent.mkdir(parents=True, exist_ok=True)
+        engine = sqa.create_engine("duckdb:///" + file)
     _sql_engine_cache["duckdb_parquet"] = engine
     path = Path(tempfile.gettempdir()) / "pdtransform" / "tests"
     file = path / f"{name}.parquet"
diff --git a/tests/util/filelock.py b/tests/util/filelock.py
new file mode 100644
index 00000000..f19c5314
--- /dev/null
+++ b/tests/util/filelock.py
@@ -0,0 +1,16 @@
+# Copyright (c) QuantCo and pydiverse contributors 2025-2025
+# SPDX-License-Identifier: BSD-3-Clause
+
+from contextlib import contextmanager
+from pathlib import Path
+
+
+@contextmanager
+def lock(lock_path: Path | str):
+    if isinstance(lock_path, str):
+        lock_path = Path(lock_path)
+    import filelock
+
+    lock_path.parent.mkdir(parents=True, exist_ok=True)
+    with filelock.FileLock(lock_path.with_suffix(".lock")):
+        yield