To fix issue 930: adding is_ordered_set_operator

cdisc_rules_engine/check_operators/dataframe_operators.py

@@ -1402,6 +1402,174 @@ def is_ordered_set(self, other_value):
     def is_not_ordered_set(self, other_value):
         return ~self.is_ordered_set(other_value)
 
+    def _coerce_sequence_value_to_int(self, value):
+        """
+        Attempts to coerce a sequence value to an integer for continuity checks.
+
+        Accepted as valid:
+        - int values
+        - float values that represent whole numbers (for example 2.0)
+        - numeric strings matching optional sign and digits (for example "-1", "03")
+
+        Rejected as invalid:
+        - null/NaN/empty values
+        - non-integer floats (for example 2.5)
+        - non-numeric strings (for example "A1", "UNK")
+
+        Returns:
+        - (True, int_value) when conversion is valid
+        - (False, None) when value cannot be used in consecutive sequence logic
+        """
+        if value is None or value == "" or pd.isna(value):
+            return False, None
+        if isinstance(value, (int, np.integer)):
+            return True, int(value)
+        if isinstance(value, (float, np.floating)):
+            if float(value).is_integer():
+                return True, int(value)
+            return False, None
+        if isinstance(value, str):
+            stripped = value.strip()
+            if re.fullmatch(r"[+-]?\d+", stripped):
+                return True, int(stripped)
+        return False, None
+
+    def _check_consecutive_partition(self, partition: pd.Series) -> pd.Series:
+        """
+        Validates consecutive sequence rules for one grouped partition of target values.
+
+        The partition is evaluated in its existing row order (same order used by
+        is_ordered_set checks). For each row:
+        - First valid numeric value initializes the running previous value.
+        - Next valid value must be equal to previous (duplicate allowed) or
+        previous + 1 (strictly consecutive step).
+        - Any null/empty/non-numeric value is marked False (strict behavior).
+        - Any numeric jump greater than 1 or decrease is marked False.
+
+        Returns:
+        - A boolean Series aligned to partition index, where each element
+        indicates whether that row satisfies the consecutive rule.
+        """
+        result = pd.Series(True, index=partition.index, dtype="bool")
+        prev_val = None
+        has_prev = False
+
+        for idx, raw in partition.items():
+            valid, current = self._coerce_sequence_value_to_int(raw)
+            if not valid:
+                result.at[idx] = False
+                continue
+
+            if not has_prev:
+                prev_val = current
+                has_prev = True
+                continue
+
+            if current == prev_val or current == prev_val + 1:
+                prev_val = current
+            else:
+                result.at[idx] = False
+                prev_val = current
+
+        return result
+
+    def _check_ordered_partition_strict(self, partition: pd.Series) -> pd.Series:
+        """
+        Validates ascending order for one grouped partition using strict numeric rules.
+
+        Rules:
+        - Values are evaluated in existing row order.
+        - Null/empty/non-numeric values are marked False.
+        - Valid numeric values must be non-decreasing (duplicates allowed).
+        """
+        result = pd.Series(True, index=partition.index, dtype="bool")
+        prev_val = None
+        has_prev = False
+
+        for idx, raw in partition.items():
+            valid, current = self._coerce_sequence_value_to_int(raw)
+            if not valid:
+                result.at[idx] = False
+                continue
+
+            if not has_prev:
+                prev_val = current
+                has_prev = True
+                continue
+
+            if current < prev_val:
+                result.at[idx] = False
+
+            prev_val = current
+
+        return result
+
+    @log_operator_execution
+    @type_operator(FIELD_DATAFRAME)
+    def is_consecutive_ordered_set(self, other_value):
+        """
+        Checks whether the values in the target column are consecutive and ordered
+        within each group defined by the comparator.
+
+        This operator extends is_ordered_set by adding continuity validation:
+        after grouping rows by comparator, target values must be in ascending order
+        and each next value must be either:
+        - the same as the previous value (duplicates allowed), or
+        - exactly previous + 1 (no skips allowed).
+
+        Strict behavior:
+        - Null, empty, or non-numeric target values are marked False.
+        - These rows are not silently ignored, so data quality issues are visible.
+
+        Parameters in other_value:
+        - target: the column containing sequence values to validate.
+        - comparator: one grouping column (string) or multiple grouping columns (list).
+
+        Example:
+        - Group USUBJID = 01 with SEQ [1, 1, 2, 3] -> all True
+        - Group USUBJID = 01 with SEQ [1, 2, 4] -> row with 4 is False (skip at 3)
+        - Group USUBJID = 01 with SEQ [1, None, 2] -> row with None is False
+        """
+        target = other_value.get("target")
+        value = other_value.get("comparator")
+
+        if not isinstance(value, (str, list)):
+            raise Exception("Comparator must be a String or list of Strings")
+        if isinstance(value, list) and not all(isinstance(v, str) for v in value):
+            raise Exception("All comparator values must be Strings")
+
+        grouping = [value] if isinstance(value, str) else value
+
+        # keep existing ordering semantics
+        data = self.value.get(grouping + [target])
+        ordered_result = (
+            data.groupby(grouping, dropna=False)[target]
+            .transform(self._check_ordered_partition_strict)
+            .sort_index()
+        )
+
+        # compute strict consecutive semantics on realized dataframe
+        consecutive_result = (
+            data.groupby(grouping, dropna=False)[target]
+            .transform(self._check_consecutive_partition)
+            .sort_index()
+        )
+
+        ordered_result = self.value.convert_to_series(ordered_result).sort_index()
+        consecutive_result = self.value.convert_to_series(consecutive_result).astype(
+            "bool"
+        )
+
+        return ordered_result & consecutive_result
+
+    @log_operator_execution
+    @type_operator(FIELD_DATAFRAME)
+    def is_not_consecutive_ordered_set(self, other_value):
+        """
+        Complement of is_consecutive_ordered_set.
+        """
+        return ~self.is_consecutive_ordered_set(other_value)
+
     @log_operator_execution
     @type_operator(FIELD_DATAFRAME)
     def has_next_corresponding_record(self, other_value: dict):

resources/schema/rule-merged/Operator.json

@@ -454,6 +454,26 @@
       "required": ["operator", "value"],
       "type": "object"
     },
+    {
+      "properties": {
+        "operator": {
+          "const": "is_consecutive_ordered_set",
+          "markdownDescription": "\nTrue if the dataset rows are in ascending order of values within `name`, grouped by `value`, and there are no skips between successive sequence values.\n\nThis extends `is_ordered_set` with continuity checking.\n\nRules:\n- Duplicates are allowed (e.g., 1, 1, 2, 3).\n- The sequence may start at any integer.\n- Null/empty/non-numeric values in `name` evaluate to false (strict behavior).\n\n```yaml\nCheck:\n  all:\n    - name: --SEQ\n      operator: is_consecutive_ordered_set\n      value: USUBJID\n```\n```yaml\nCheck:\n  all:\n    - name: --SEQ\n      operator: is_consecutive_ordered_set\n      value:\n        - USUBJID\n        - \"--TESTCD\"\n```\n"
+        }
+      },
+      "required": ["operator", "value"],
+      "type": "object"
+    },
+    {
+      "properties": {
+        "operator": {
+          "const": "is_not_consecutive_ordered_set",
+          "markdownDescription": "\nComplement of `is_consecutive_ordered_set`\n"
+        }
+      },
+      "required": ["operator", "value"],
+      "type": "object"
+    },
     {
       "properties": {
         "operator": {

resources/schema/rule/Operator.json

@@ -246,6 +246,16 @@
       "required": ["operator", "value"],
       "type": "object"
     },
+    {
+      "properties": { "operator": { "const": "is_consecutive_ordered_set" } },
+      "required": ["operator", "value"],
+      "type": "object"
+    },
+    {
+      "properties": { "operator": { "const": "is_not_consecutive_ordered_set" } },
+      "required": ["operator", "value"],
+      "type": "object"
+    },
     {
       "properties": { "operator": { "const": "is_unique_relationship" } },
       "required": ["operator", "value"],

resources/schema/rule/Operator.md

@@ -1113,6 +1113,38 @@ Check:
         - "--TESTCD"
 ```
 
+### is_consecutive_ordered_set
+
+True if the dataset rows are in ascending order of values within `name`, grouped by `value`, and there are no skips between successive sequence values.
+
+This extends `is_ordered_set` with continuity checking.
+
+Rules:
+- Duplicates are allowed (e.g., 1, 1, 2, 3).
+- The sequence may start at any integer.
+- Null/empty/non-numeric values in `name` evaluate to false (strict behavior).
+
+```yaml
+Check:
+  all:
+    - name: --SEQ
+      operator: is_consecutive_ordered_set
+      value: USUBJID
+```
+```yaml
+Check:
+  all:
+    - name: --SEQ
+      operator: is_consecutive_ordered_set
+      value:
+        - USUBJID
+        - "--TESTCD"
+```
+
+### is_not_consecutive_ordered_set
+
+Complement of `is_consecutive_ordered_set`
+
 ### is_ordered_by
 
 True if the dataset rows are ordered by the values within `name`, given the ordering specified by `order`

tests/unit/test_check_operators/test_value_set_checks.py

@@ -218,6 +218,97 @@ def test_is_ordered_set_multiple_comparators():
     )
 
 
+def test_is_consecutive_ordered_set():
+    data = {
+        "GROUP": ["A", "A", "A", "B", "B", "B", "C", "C"],
+        "VALUE": [1, 2, 3, 1, 2, 4, 1, 1],
+    }
+    df = PandasDataset.from_dict(data)
+    result = DataframeType({"value": df}).is_consecutive_ordered_set(
+        {"target": "VALUE", "comparator": "GROUP"}
+    )
+    pd.testing.assert_series_equal(
+        result,
+        pd.Series([True, True, True, True, True, False, True, True]),
+        check_names=False,
+    )
+
+
+@pytest.mark.parametrize("dataset_type", [PandasDataset, DaskDataset])
+def test_is_consecutive_ordered_set_strict_invalid_values(dataset_type):
+    data = {
+        "GROUP": ["A", "A", "A", "A", "A"],
+        "VALUE": [1, None, 2, "X", 3],
+    }
+    df = dataset_type.from_dict(data)
+    result = DataframeType({"value": df}).is_consecutive_ordered_set(
+        {"target": "VALUE", "comparator": "GROUP"}
+    )
+    pd.testing.assert_series_equal(
+        result, pd.Series([True, False, True, False, True]), check_names=False
+    )
+
+
+@pytest.mark.parametrize("dataset_type", [PandasDataset, DaskDataset])
+def test_is_not_consecutive_ordered_set(dataset_type):
+    data = {
+        "GROUP": ["A", "A", "A"],
+        "VALUE": [1, 2, 4],
+    }
+    df = dataset_type.from_dict(data)
+    result = DataframeType({"value": df}).is_not_consecutive_ordered_set(
+        {"target": "VALUE", "comparator": "GROUP"}
+    )
+    pd.testing.assert_series_equal(
+        result, pd.Series([False, False, True]), check_names=False
+    )
+
+
+@pytest.mark.parametrize("dataset_type", [PandasDataset, DaskDataset])
+def test_is_consecutive_ordered_set_unordered_fails(dataset_type):
+    data = {
+        "GROUP": ["A", "A", "A"],
+        "VALUE": [1, 3, 2],
+    }
+    df = dataset_type.from_dict(data)
+    result = DataframeType({"value": df}).is_consecutive_ordered_set(
+        {"target": "VALUE", "comparator": "GROUP"}
+    )
+    pd.testing.assert_series_equal(
+        result, pd.Series([True, False, False]), check_names=False
+    )
+
+
+@pytest.mark.parametrize("dataset_type", [PandasDataset, DaskDataset])
+def test_is_consecutive_ordered_set_accepts_numeric_strings(dataset_type):
+    data = {
+        "GROUP": ["A", "A", "A"],
+        "VALUE": [1, "2", 3],
+    }
+    df = dataset_type.from_dict(data)
+    result = DataframeType({"value": df}).is_consecutive_ordered_set(
+        {"target": "VALUE", "comparator": "GROUP"}
+    )
+    pd.testing.assert_series_equal(
+        result, pd.Series([True, True, True]), check_names=False
+    )
+
+
+@pytest.mark.parametrize("dataset_type", [PandasDataset, DaskDataset])
+def test_is_not_consecutive_ordered_set_accepts_numeric_strings(dataset_type):
+    data = {
+        "GROUP": ["A", "A", "A"],
+        "VALUE": [1, "2", 3],
+    }
+    df = dataset_type.from_dict(data)
+    result = DataframeType({"value": df}).is_not_consecutive_ordered_set(
+        {"target": "VALUE", "comparator": "GROUP"}
+    )
+    pd.testing.assert_series_equal(
+        result, pd.Series([False, False, False]), check_names=False
+    )
+
+
 @pytest.mark.parametrize(
     "target, comparator, dataset_type, expected_result",
     [