tablevalidator-databricks 0.1.1__tar.gz → 0.1.3__tar.gz

{tablevalidator_databricks-0.1.1 → tablevalidator_databricks-0.1.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: tablevalidator-databricks
-Version: 0.1.1
+Version: 0.1.3
 Summary: Generates a widget-driven Databricks notebook UI on top of the table-validator package - no code required to run a comparison.
 License: MIT
 Requires-Python: >=3.9
@@ -45,11 +45,12 @@ Writes `./TableValidator.py` - a ready-to-use Databricks notebook. Options:
 tablevalidator-databricks init --output MyValidation.py --mode full
 ```
 
-- `--mode basic` - only the check-group multiselect and table pickers.
+- `--mode basic` - only the check-group Yes/No dropdowns and table pickers.
 - `--mode full` (default) - also shows optional filter widgets (only/
-  ignore columns, row filter, primary key).
-- `--mode schema` - pre-selects only the "Catalog & Schema" check group,
-  for a lightweight schema-shape-only comparison.
+  ignore columns, row filter, primary key). Every check defaults to "Yes".
+- `--mode schema` - same widgets as `basic`, but Column/Row default to
+  "No" - a lightweight schema-shape-only comparison out of the box (still
+  changeable per-run).
 
 Then, in your Databricks workspace: **Workspace -> Import**, format
 "Source", and select the generated file. Databricks recognizes the
@@ -62,7 +63,7 @@ markers and opens it as a real notebook.
 2. Queries Unity Catalog (`SHOW CATALOGS` / `SHOW SCHEMAS IN ...` /
    `SHOW TABLES IN ...` via the notebook's own ambient Spark session) to
    populate Source/Target Catalog, Schema, and Table dropdowns.
-3. Shows a "Validation Checks" multiselect (Catalog & Schema / Column /
+3. Shows a Yes/No dropdown per check group (Catalog & Schema / Column /
    Row - the three independently-selectable check groups the engine
    actually supports) plus, in `full` mode, optional text widgets for
    column filtering, a row filter, and a primary key.
@@ -86,6 +87,14 @@ isn't showing up (a permissions lag, sort order, anything). It's used
 as-is, not checked against the live catalog listing. Leave every override
 blank to just use the dropdowns as normal.
 
+**Validate every table in a schema, not just one**: pick
+"(all tables in schema)" as the Table dropdown (Source and/or Target
+independently) to compare every identically-named table in that schema in
+one run - the same schema-wide sweep `validate_tables()` itself supports.
+A primary key can't be set in this mode (leave the Primary Key widget
+blank). Comparing every schema in a catalog isn't supported yet - Schema
+must be a specific choice, even when Table is left as "(all ... in ...)".
+
 ## Why a separate package
 
 `table-validator` stays focused on the actual comparison engine (CLI +

{tablevalidator_databricks-0.1.1 → tablevalidator_databricks-0.1.3}/README.md

@@ -31,11 +31,12 @@ Writes `./TableValidator.py` - a ready-to-use Databricks notebook. Options:
 tablevalidator-databricks init --output MyValidation.py --mode full
 ```
 
-- `--mode basic` - only the check-group multiselect and table pickers.
+- `--mode basic` - only the check-group Yes/No dropdowns and table pickers.
 - `--mode full` (default) - also shows optional filter widgets (only/
-  ignore columns, row filter, primary key).
-- `--mode schema` - pre-selects only the "Catalog & Schema" check group,
-  for a lightweight schema-shape-only comparison.
+  ignore columns, row filter, primary key). Every check defaults to "Yes".
+- `--mode schema` - same widgets as `basic`, but Column/Row default to
+  "No" - a lightweight schema-shape-only comparison out of the box (still
+  changeable per-run).
 
 Then, in your Databricks workspace: **Workspace -> Import**, format
 "Source", and select the generated file. Databricks recognizes the
@@ -48,7 +49,7 @@ markers and opens it as a real notebook.
 2. Queries Unity Catalog (`SHOW CATALOGS` / `SHOW SCHEMAS IN ...` /
    `SHOW TABLES IN ...` via the notebook's own ambient Spark session) to
    populate Source/Target Catalog, Schema, and Table dropdowns.
-3. Shows a "Validation Checks" multiselect (Catalog & Schema / Column /
+3. Shows a Yes/No dropdown per check group (Catalog & Schema / Column /
    Row - the three independently-selectable check groups the engine
    actually supports) plus, in `full` mode, optional text widgets for
    column filtering, a row filter, and a primary key.
@@ -72,6 +73,14 @@ isn't showing up (a permissions lag, sort order, anything). It's used
 as-is, not checked against the live catalog listing. Leave every override
 blank to just use the dropdowns as normal.
 
+**Validate every table in a schema, not just one**: pick
+"(all tables in schema)" as the Table dropdown (Source and/or Target
+independently) to compare every identically-named table in that schema in
+one run - the same schema-wide sweep `validate_tables()` itself supports.
+A primary key can't be set in this mode (leave the Primary Key widget
+blank). Comparing every schema in a catalog isn't supported yet - Schema
+must be a specific choice, even when Table is left as "(all ... in ...)".
+
 ## Why a separate package
 
 `table-validator` stays focused on the actual comparison engine (CLI +

{tablevalidator_databricks-0.1.1 → tablevalidator_databricks-0.1.3}/tablevalidator_databricks/templates/widget_notebook.py.tmpl

@@ -22,6 +22,12 @@
 # MAGIC something isn't showing up (a permissions lag, a name that doesn't sort
 # MAGIC into view, etc.). Leave every override blank to just use the dropdowns
 # MAGIC as normal.
+# MAGIC
+# MAGIC **Want to validate every table in a schema instead of just one?** Pick
+# MAGIC "(all tables in schema)" as the Table (both Source and Target) - every
+# MAGIC identically-named table in that schema is compared in one run. A
+# MAGIC primary key can't be set in this mode (a single key can't apply to
+# MAGIC every table) - leave the Primary Key widget blank if you use this.
 
 # COMMAND ----------
 
@@ -66,6 +72,30 @@ def _current_or_first(options, current):
         return current
     return options[0] if options else ""
 
+
+def _current_or_first_real(options, current, sentinel):
+    """Like _current_or_first, but when falling back (no valid current
+    selection), prefers the first REAL option over `sentinel` if one
+    exists - so a brand-new notebook defaults to an actual Schema/Table
+    rather than immediately landing on "(all ... in ...)". The sentinel
+    stays fully selectable (and sticky once picked) via the current-
+    value check above; this only changes what a first-time default
+    lands on."""
+    if current in options:
+        return current
+    real_options = [o for o in options if o != sentinel]
+    return real_options[0] if real_options else (options[0] if options else "")
+
+
+# Sentinel choices meaning "leave this blank" - selecting one of these
+# skips the level(s) below it and triggers validate_tables()'s own
+# schema-wide sweep mode (comparing every identically-named table)
+# instead of a single table pair. Real Databricks catalog/schema/table
+# names can't contain parentheses, so these can never collide with a
+# genuine name.
+_ALL_SCHEMAS = "(all schemas in catalog)"
+_ALL_TABLES = "(all tables in schema)"
+
 # COMMAND ----------
 
 # MAGIC %md ### Create / refresh widgets
@@ -103,18 +133,34 @@ def _build_cascading_widgets(prefix: str, label_prefix: str) -> None:
     now, not a stale snapshot from an earlier run. Also creates a
     "(manual override)" text widget next to each dropdown, for a name
     that isn't showing up in the dropdown for any reason - see
-    _effective_value() below for how the override takes priority."""
+    _effective_value() below for how the override takes priority.
+
+    Schema and Table each also get an "(all ... in ...)" sentinel choice
+    as their first option, for a schema-wide sweep (leave Table blank to
+    compare every identically-named table in the schema) or a catalog-
+    wide sweep (leave Schema, and therefore Table too, blank to compare
+    every identically-named table across every schema) - see
+    _ALL_SCHEMAS/_ALL_TABLES above and how the "Run validation" cell
+    below builds validate_tables()'s "catalog.schema" (no table) /
+    "catalog" (no schema) sweep-mode strings from this choice."""
     catalog_current = _current_or_first(catalogs, _effective_value(f"{prefix}_catalog"))
-    schemas = _list_schemas(catalog_current)
-    schema_current = _current_or_first(schemas, _effective_value(f"{prefix}_schema"))
-    tables = _list_tables(catalog_current, schema_current)
-    table_current = _current_or_first(tables, _effective_value(f"{prefix}_table"))
+
+    schema_choices = [_ALL_SCHEMAS] + _list_schemas(catalog_current)
+    schema_current = _current_or_first_real(schema_choices, _effective_value(f"{prefix}_schema"), _ALL_SCHEMAS)
+    sweeping_schema = schema_current == _ALL_SCHEMAS
+
+    if sweeping_schema:
+        table_choices = [_ALL_TABLES]
+        table_current = _ALL_TABLES
+    else:
+        table_choices = [_ALL_TABLES] + _list_tables(catalog_current, schema_current)
+        table_current = _current_or_first_real(table_choices, _effective_value(f"{prefix}_table"), _ALL_TABLES)
 
     dbutils.widgets.dropdown(f"{prefix}_catalog", catalog_current, catalogs or [""], f"{label_prefix} Catalog")
     dbutils.widgets.text(f"{prefix}_catalog_override", "", f"{label_prefix} Catalog (manual override)")
-    dbutils.widgets.dropdown(f"{prefix}_schema", schema_current, schemas or [""], f"{label_prefix} Schema")
+    dbutils.widgets.dropdown(f"{prefix}_schema", schema_current, schema_choices, f"{label_prefix} Schema")
     dbutils.widgets.text(f"{prefix}_schema_override", "", f"{label_prefix} Schema (manual override)")
-    dbutils.widgets.dropdown(f"{prefix}_table", table_current, tables or [""], f"{label_prefix} Table")
+    dbutils.widgets.dropdown(f"{prefix}_table", table_current, table_choices, f"{label_prefix} Table")
     dbutils.widgets.text(f"{prefix}_table_override", "", f"{label_prefix} Table (manual override)")
 
 
@@ -134,11 +180,20 @@ print(
 
 # COMMAND ----------
 
-_default_checks = ["Catalog & Schema", "Column", "Row"] if MODE != "schema" else ["Catalog & Schema"]
-dbutils.widgets.multiselect(
-    "checks", ",".join(_default_checks),
-    ["Catalog & Schema", "Column", "Row"], "Validation Checks",
-)
+# One Yes/No dropdown per check group, instead of a single multiselect -
+# dbutils.widgets.multiselect enforces an internal "selection sequence"
+# validation against its own default that was observed to reject even a
+# same-length, same-items default on some Databricks runtimes/existing-
+# widget states (DefaultValueNotInChoicesList), with no reliable
+# workaround. Three independent dropdowns have no such cross-widget
+# validation and are the more robust primitive for this.
+_YES_NO = ["Yes", "No"]
+_schema_default = "Yes"
+_column_default = "No" if MODE == "schema" else "Yes"
+_row_default = "No" if MODE == "schema" else "Yes"
+dbutils.widgets.dropdown("check_catalog_schema", _schema_default, _YES_NO, "Run Catalog & Schema checks?")
+dbutils.widgets.dropdown("check_column", _column_default, _YES_NO, "Run Column checks?")
+dbutils.widgets.dropdown("check_row", _row_default, _YES_NO, "Run Row checks?")
 
 if SHOW_OPTIONAL_WIDGETS:
     dbutils.widgets.text("only_columns", "", "Only compare these columns (comma-separated, optional)")
@@ -161,14 +216,15 @@ def _validate_current_selection(prefix: str, label: str) -> None:
     manual override set is trusted as typed and skipped here - the whole
     point of the override is to name something the live catalog listing
     doesn't (currently) show, so it can never pass this check by
-    definition."""
+    definition. _ALL_SCHEMAS/_ALL_TABLES (leave-blank sweep sentinels)
+    are always valid and skip the corresponding check."""
     catalog = _effective_value(f"{prefix}_catalog")
     schema_overridden = bool(_get_widget(f"{prefix}_schema_override").strip())
     table_overridden = bool(_get_widget(f"{prefix}_table_override").strip())
     schema = _effective_value(f"{prefix}_schema")
     table = _effective_value(f"{prefix}_table")
 
-    if not schema_overridden:
+    if not schema_overridden and schema != _ALL_SCHEMAS:
         live_schemas = _list_schemas(catalog)
         if schema not in live_schemas:
             raise ValueError(
@@ -181,7 +237,7 @@ def _validate_current_selection(prefix: str, label: str) -> None:
                 "the dropdown, type it into the Schema manual override "
                 "widget instead.)"
             )
-    if not table_overridden:
+    if not table_overridden and schema != _ALL_SCHEMAS and table != _ALL_TABLES:
         live_tables = _list_tables(catalog, schema)
         if table not in live_tables:
             raise ValueError(
@@ -198,18 +254,40 @@ def _validate_current_selection(prefix: str, label: str) -> None:
 _validate_current_selection("source", "Source")
 _validate_current_selection("target", "Target")
 
-source = f"{_effective_value('source_catalog')}.{_effective_value('source_schema')}.{_effective_value('source_table')}"
-target = f"{_effective_value('target_catalog')}.{_effective_value('target_schema')}.{_effective_value('target_table')}"
 
-_CHECK_MAP = {
-    "Catalog & Schema": {ValidationType.CATALOG, ValidationType.SCHEMA},
-    "Column": {ValidationType.COLUMN},
-    "Row": {ValidationType.ROW},
-}
-selected_checks = [c for c in dbutils.widgets.get("checks").split(",") if c]
+def _source_or_target_string(prefix: str) -> str:
+    """Builds the "catalog.schema.table" (single table) or
+    "catalog.schema" (schema-wide sweep, Schema chosen but Table left as
+    _ALL_TABLES) string validate_tables() expects. Catalog-wide sweeps
+    (Schema also left blank) aren't supported by validate_tables() today
+    - _ALL_SCHEMAS is caught with a clear error below instead of being
+    silently passed through as a literal, broken source/target string."""
+    catalog = _effective_value(f"{prefix}_catalog")
+    schema = _effective_value(f"{prefix}_schema")
+    table = _effective_value(f"{prefix}_table")
+
+    if schema == _ALL_SCHEMAS:
+        raise ValueError(
+            f"{prefix.capitalize()} Schema is set to {_ALL_SCHEMAS!r} - "
+            "comparing every schema in a catalog isn't supported yet. "
+            "Pick a specific Schema (Table can still be left as "
+            f"{_ALL_TABLES!r} to sweep every table within it)."
+        )
+    if table == _ALL_TABLES:
+        return f"{catalog}.{schema}"
+    return f"{catalog}.{schema}.{table}"
+
+
+source = _source_or_target_string("source")
+target = _source_or_target_string("target")
+
 enabled_validations = set()
-for check in selected_checks:
-    enabled_validations |= _CHECK_MAP.get(check, set())
+if dbutils.widgets.get("check_catalog_schema") == "Yes":
+    enabled_validations |= {ValidationType.CATALOG, ValidationType.SCHEMA}
+if dbutils.widgets.get("check_column") == "Yes":
+    enabled_validations |= {ValidationType.COLUMN}
+if dbutils.widgets.get("check_row") == "Yes":
+    enabled_validations |= {ValidationType.ROW}
 if not enabled_validations:
     enabled_validations = None  # fall back to validate_tables()'s own "run everything" default
 

{tablevalidator_databricks-0.1.1 → tablevalidator_databricks-0.1.3}/tablevalidator_databricks.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: tablevalidator-databricks
-Version: 0.1.1
+Version: 0.1.3
 Summary: Generates a widget-driven Databricks notebook UI on top of the table-validator package - no code required to run a comparison.
 License: MIT
 Requires-Python: >=3.9
@@ -45,11 +45,12 @@ Writes `./TableValidator.py` - a ready-to-use Databricks notebook. Options:
 tablevalidator-databricks init --output MyValidation.py --mode full
 ```
 
-- `--mode basic` - only the check-group multiselect and table pickers.
+- `--mode basic` - only the check-group Yes/No dropdowns and table pickers.
 - `--mode full` (default) - also shows optional filter widgets (only/
-  ignore columns, row filter, primary key).
-- `--mode schema` - pre-selects only the "Catalog & Schema" check group,
-  for a lightweight schema-shape-only comparison.
+  ignore columns, row filter, primary key). Every check defaults to "Yes".
+- `--mode schema` - same widgets as `basic`, but Column/Row default to
+  "No" - a lightweight schema-shape-only comparison out of the box (still
+  changeable per-run).
 
 Then, in your Databricks workspace: **Workspace -> Import**, format
 "Source", and select the generated file. Databricks recognizes the
@@ -62,7 +63,7 @@ markers and opens it as a real notebook.
 2. Queries Unity Catalog (`SHOW CATALOGS` / `SHOW SCHEMAS IN ...` /
    `SHOW TABLES IN ...` via the notebook's own ambient Spark session) to
    populate Source/Target Catalog, Schema, and Table dropdowns.
-3. Shows a "Validation Checks" multiselect (Catalog & Schema / Column /
+3. Shows a Yes/No dropdown per check group (Catalog & Schema / Column /
    Row - the three independently-selectable check groups the engine
    actually supports) plus, in `full` mode, optional text widgets for
    column filtering, a row filter, and a primary key.
@@ -86,6 +87,14 @@ isn't showing up (a permissions lag, sort order, anything). It's used
 as-is, not checked against the live catalog listing. Leave every override
 blank to just use the dropdowns as normal.
 
+**Validate every table in a schema, not just one**: pick
+"(all tables in schema)" as the Table dropdown (Source and/or Target
+independently) to compare every identically-named table in that schema in
+one run - the same schema-wide sweep `validate_tables()` itself supports.
+A primary key can't be set in this mode (leave the Primary Key widget
+blank). Comparing every schema in a catalog isn't supported yet - Schema
+must be a specific choice, even when Table is left as "(all ... in ...)".
+
 ## Why a separate package
 
 `table-validator` stays focused on the actual comparison engine (CLI +

{tablevalidator_databricks-0.1.1 → tablevalidator_databricks-0.1.3}/tablevalidator_databricks.egg-info/scm_version.json

@@ -1,7 +1,7 @@
 {
-  "tag": "0.1.1",
+  "tag": "0.1.3",
   "distance": 0,
-  "node": "g1f85a99",
+  "node": "gd05d199",
   "dirty": false,
   "branch": "HEAD",
   "node_date": "2026-09-08"

{tablevalidator_databricks-0.1.1 → tablevalidator_databricks-0.1.3}/tests/test_generated_notebook_execution.py

@@ -15,6 +15,7 @@ from unittest.mock import MagicMock
 
 import pytest
 
+from table_validator.config.schema import ValidationType
 from tablevalidator_databricks.generator.generator import generate_notebook
 
 
@@ -29,9 +30,15 @@ def _code_from_notebook(path: Path) -> str:
     return "\n".join(code_lines)
 
 
+class _DefaultValueNotInChoicesList(Exception):
+    pass
+
+
 class _FakeWidgets:
     def __init__(self, initial=None):
         self._values = dict(initial or {})
+        self.dropdown_choices = {}
+        self.multiselect_choices = {}
 
     def get(self, name):
         return self._values.get(name, "")
@@ -43,9 +50,25 @@ class _FakeWidgets:
         # Real dbutils.widgets.dropdown does NOT overwrite an existing
         # value just because the cell re-ran with a different `default` -
         # it only sets it the first time. Match that.
+        if default not in choices:
+            raise _DefaultValueNotInChoicesList(f"{name}: {default!r} not in {choices!r}")
+        self.dropdown_choices[name] = choices
         self._values.setdefault(name, default)
 
     def multiselect(self, name, default, choices, label):
+        # Real Databricks dbutils.widgets.multiselect requires the
+        # default's comma-separated items to all be present in choices -
+        # observed in production to reject a narrower/reordered subset
+        # with "DefaultValueNotInChoicesList: Selection sequence must
+        # include <choices>" even when every individual item IS a valid
+        # choice. Replicate that constraint here so this suite would have
+        # caught the real bug (a "schema"-mode default of just
+        # "Catalog & Schema" against choices=["Catalog & Schema",
+        # "Column", "Row"]) before it ever reached a live workspace.
+        default_items = default.split(",") if default else []
+        if any(item not in choices for item in default_items):
+            raise _DefaultValueNotInChoicesList(f"{name}: {default!r} not a subset of {choices!r}")
+        self.multiselect_choices[name] = choices
         self._values.setdefault(name, default)
 
 
@@ -229,3 +252,166 @@ def test_catalog_override_drives_schema_and_table_lookup(notebook_path):
     )
 
     assert captured["source"] == "hidden_catalog.sales.orders"
+
+
+# ---------------------------------------------------------------------------
+# Validation-checks selection - three independent Yes/No dropdowns
+# (check_catalog_schema / check_column / check_row), replacing an earlier
+# single dbutils.widgets.multiselect that was observed in production to
+# raise DefaultValueNotInChoicesList even with a default that appeared to
+# satisfy multiselect's own documented constraints - three independent
+# dropdowns have no equivalent cross-widget validation.
+# ---------------------------------------------------------------------------
+def test_full_and_basic_mode_default_all_checks_to_yes(tmp_path: Path) -> None:
+    for mode in ("full", "basic"):
+        path = tmp_path / f"{mode}.py"
+        generate_notebook(path, mode)
+        spark = _fake_spark(
+            catalogs=["main"], schemas_by_catalog={"main": ["sales"]},
+            tables_by_schema={("main", "sales"): ["orders"]},
+        )
+        captured = {}
+
+        def fake_validate_tables(source, target, **kwargs):
+            captured["enabled_validations"] = kwargs.get("enabled_validations")
+            m = MagicMock()
+            m.__str__ = MagicMock(return_value="Overall status: PASS")
+            return m
+
+        _run_notebook(path, widget_values={}, spark=spark, validate_tables_mock=fake_validate_tables)
+
+        assert captured["enabled_validations"] == {
+            ValidationType.CATALOG, ValidationType.SCHEMA,
+            ValidationType.COLUMN, ValidationType.ROW,
+        }, f"mode={mode}"
+
+
+def test_schema_mode_defaults_column_and_row_checks_to_no(tmp_path: Path) -> None:
+    path = tmp_path / "schema.py"
+    generate_notebook(path, "schema")
+    spark = _fake_spark(
+        catalogs=["main"], schemas_by_catalog={"main": ["sales"]},
+        tables_by_schema={("main", "sales"): ["orders"]},
+    )
+    captured = {}
+
+    def fake_validate_tables(source, target, **kwargs):
+        captured["enabled_validations"] = kwargs.get("enabled_validations")
+        m = MagicMock()
+        m.__str__ = MagicMock(return_value="Overall status: PASS")
+        return m
+
+    _run_notebook(path, widget_values={}, spark=spark, validate_tables_mock=fake_validate_tables)
+
+    assert captured["enabled_validations"] == {ValidationType.CATALOG, ValidationType.SCHEMA}
+
+
+def test_deselecting_every_check_falls_back_to_run_everything(notebook_path):
+    """All three Yes/No dropdowns set to "No" must fall back to
+    validate_tables()'s own "run everything" default (enabled_validations=
+    None), not silently run nothing."""
+    spark = _fake_spark(
+        catalogs=["main"], schemas_by_catalog={"main": ["sales"]},
+        tables_by_schema={("main", "sales"): ["orders"]},
+    )
+    captured = {}
+
+    def fake_validate_tables(source, target, **kwargs):
+        captured["enabled_validations"] = kwargs.get("enabled_validations")
+        m = MagicMock()
+        m.__str__ = MagicMock(return_value="Overall status: PASS")
+        return m
+
+    _run_notebook(
+        notebook_path,
+        widget_values={
+            "check_catalog_schema": "No", "check_column": "No", "check_row": "No",
+        },
+        spark=spark,
+        validate_tables_mock=fake_validate_tables,
+    )
+
+    assert captured["enabled_validations"] is None
+
+
+# ---------------------------------------------------------------------------
+# Schema-wide sweep mode - leaving Table as "(all tables in schema)".
+# ---------------------------------------------------------------------------
+def test_default_selection_lands_on_a_real_table_not_the_sweep_sentinel(notebook_path):
+    """A brand-new notebook (no prior widget values) must default to a
+    real Catalog/Schema/Table, not silently land on the "(all ... in
+    ...)" sweep sentinel and immediately error - the sentinel must be
+    reachable but never the unprompted default."""
+    spark = _fake_spark(
+        catalogs=["main"],
+        schemas_by_catalog={"main": ["sales"]},
+        tables_by_schema={("main", "sales"): ["orders"]},
+    )
+    captured = {}
+
+    def fake_validate_tables(source, target, **kwargs):
+        captured["source"] = source
+        captured["target"] = target
+        m = MagicMock()
+        m.__str__ = MagicMock(return_value="Overall status: PASS")
+        return m
+
+    _run_notebook(notebook_path, widget_values={}, spark=spark, validate_tables_mock=fake_validate_tables)
+
+    assert captured["source"] == "main.sales.orders"
+
+
+def test_table_left_as_sweep_sentinel_builds_two_part_source_string(notebook_path):
+    """Selecting "(all tables in schema)" for Table must produce a
+    "catalog.schema" (no table) string - validate_tables()'s own
+    schema-wide sweep trigger - comparing every identically-named table."""
+    spark = _fake_spark(
+        catalogs=["main"],
+        schemas_by_catalog={"main": ["bronze"], },
+        tables_by_schema={("main", "bronze"): ["orders", "customers"]},
+    )
+    captured = {}
+
+    def fake_validate_tables(source, target, **kwargs):
+        captured["source"] = source
+        captured["target"] = target
+        captured["kwargs"] = kwargs
+        m = MagicMock()
+        m.__str__ = MagicMock(return_value="Overall status: PASS")
+        return m
+
+    _run_notebook(
+        notebook_path,
+        widget_values={
+            "source_catalog": "main", "source_schema": "bronze", "source_table": "(all tables in schema)",
+            "target_catalog": "main", "target_schema": "bronze", "target_table": "(all tables in schema)",
+        },
+        spark=spark,
+        validate_tables_mock=fake_validate_tables,
+    )
+
+    assert captured["source"] == "main.bronze"
+    assert captured["target"] == "main.bronze"
+
+
+def test_catalog_wide_sweep_raises_clear_error(notebook_path):
+    """Schema left as "(all schemas in catalog)" isn't supported by
+    validate_tables() today - must fail with a clear, actionable error
+    rather than building a broken "catalog" (no schema) string."""
+    spark = _fake_spark(
+        catalogs=["main"],
+        schemas_by_catalog={"main": ["sales"]},
+        tables_by_schema={("main", "sales"): ["orders"]},
+    )
+
+    with pytest.raises(ValueError, match="all schemas in catalog"):
+        _run_notebook(
+            notebook_path,
+            widget_values={
+                "source_catalog": "main", "source_schema": "(all schemas in catalog)",
+                "source_table": "(all tables in schema)",
+                "target_catalog": "main", "target_schema": "sales", "target_table": "orders",
+            },
+            spark=spark,
+            validate_tables_mock=MagicMock(),
+        )