update for more valid combos

Author: sungchun12

docs/guides/linter.md

@@ -75,7 +75,7 @@ Here are all of SQLMesh's built-in linting rules:
 | `noselectstar`               | Stylistic   | The query's top-level selection may not be `SELECT *`, even if SQLMesh can expand the `SELECT *` into individual columns |
 | `nomissingaudits`             | Governance  | SQLMesh did not find any `audits` in the model's configuration to test data quality.                                                 |
 | `nomissingexternalmodels`   | Governance  | All external models must be registered in the external_models.yaml file                                                |
-| `cronintervalalignment`             | Governance  | Upstream model has a cron expression with longer intervals than downstream model. Example:  step_1(`@weekly`) -> step_2(`@daily`) -> step_3(`*/5 * * * *`). step_2 and step_3 are anchored to step_1's cron and will run on the same schedule as step_1. The fix is to align the schedules where a downstream model's cron is the same or has a longer cron interval than an upstream model's.                                         |
+| `cronintervalalignment`             | Governance  | Upstream model has a cron expression with longer intervals than downstream model. Example:  step_1(`@daily`) -> step_2(`@hourly`) -> step_3(`*/5 * * * *`). step_2 and step_3 will not behave as intended because their temporal grains are smaller than step_1's. step_1 will backfill all of `2020-01-01 - 2020-01-01`. step_2 will backfill every hour interval starting from `2020-01-02 00:00:00 - 2020-01-02 00:01:00`. However, it will wastefully backfill empty intervals given step_1 does not contain data for this interval. This can exacerbate scheduling inefficiencies. The fix is to align the schedules where a downstream model's cron is the same or has a longer cron interval than an upstream model's.                                          |
 
 ### User-defined rules
 

sqlmesh/core/linter/rules/builtin.py

@@ -277,11 +277,13 @@ def create_fix(self, model_name: str) -> t.Optional[Fix]:
 class CronIntervalAlignment(Rule):
     """Upstream model has a cron expression with longer intervals than downstream model."""
 
-    def check_model(self, model: Model) -> t.Optional[RuleViolation]:
+    def check_model(self, model: Model) -> t.Optional[t.List[RuleViolation]]:
         placeholder_start_date = "2020-01-01 10:00:00"
 
         this_model_cron_next = model.cron_next(placeholder_start_date)
 
+        violations = []
+        has_valid_upstream = False
         for upstream_model_name in model.depends_on:
             upstream_model = self.context.get_model(upstream_model_name)
 
@@ -297,11 +299,19 @@ def check_model(self, model: Model) -> t.Optional[RuleViolation]:
             upstream_model_cron_next = upstream_model.cron_next(placeholder_start_date)
 
             if upstream_model_cron_next > this_model_cron_next:
-                return self.violation(
-                    f"Upstream model {upstream_model_name} has longer cron interval ({upstream_model.cron}) "
-                    f"than this model ({model.cron})"
+                violations.append(
+                    RuleViolation(
+                        rule=self,
+                        violation_msg=f"Upstream model {upstream_model_name} has longer cron interval ({upstream_model.cron}) "
+                        f"than this model ({model.cron})",
+                    )
                 )
-        return None
+            elif upstream_model_cron_next <= this_model_cron_next:
+                has_valid_upstream = True
+                break
+        if has_valid_upstream:
+            return None
+        return violations
 
 
 BUILTIN_RULES = RuleSet(subclasses(__name__, Rule, (Rule,)))