SQL rendering tests for no-metric queries with custom granularity (#1411

)

Tests for no-metric queries with custom granularities. Does not include
tests with filters (coming later). If you have suggestions for
additional no-metric scenarios that should be tested, please let me
know!
Note that there is a bug in the optimization for one of these tests. I
have a task to fix this up stack.

tests_metricflow/query_rendering/test_custom_granularity.py

@@ -14,6 +14,7 @@
 from metricflow_semantics.specs.query_spec import MetricFlowQuerySpec
 from metricflow_semantics.specs.time_dimension_spec import TimeDimensionSpec
 from metricflow_semantics.test_helpers.config_helpers import MetricFlowTestConfiguration
+from metricflow_semantics.test_helpers.metric_time_dimension import MTD_SPEC_DAY
 from metricflow_semantics.time.granularity import ExpandedTimeGranularity
 
 from metricflow.dataflow.builder.dataflow_plan_builder import DataflowPlanBuilder
@@ -161,3 +162,79 @@ def test_metric_custom_granularity_joined_to_non_default_grain(  # noqa: D103
         dataflow_plan_builder=dataflow_plan_builder,
         query_spec=query_spec,
     )
+
+
+@pytest.mark.sql_engine_snapshot
+def test_no_metric_custom_granularity_non_metric_time(  # noqa: D103
+    request: FixtureRequest,
+    mf_test_configuration: MetricFlowTestConfiguration,
+    dataflow_plan_builder: DataflowPlanBuilder,
+    dataflow_to_sql_converter: DataflowToSqlQueryPlanConverter,
+    sql_client: SqlClient,
+) -> None:
+    query_spec = MetricFlowQuerySpec(
+        time_dimension_specs=(normal_time_dim_with_custom_grain1,),
+    )
+
+    render_and_check(
+        request=request,
+        mf_test_configuration=mf_test_configuration,
+        dataflow_to_sql_converter=dataflow_to_sql_converter,
+        sql_client=sql_client,
+        dataflow_plan_builder=dataflow_plan_builder,
+        query_spec=query_spec,
+    )
+
+
+@pytest.mark.sql_engine_snapshot
+def test_no_metric_custom_granularity_joined_to_non_default_grain(  # noqa: D103
+    request: FixtureRequest,
+    mf_test_configuration: MetricFlowTestConfiguration,
+    dataflow_plan_builder: DataflowPlanBuilder,
+    dataflow_to_sql_converter: DataflowToSqlQueryPlanConverter,
+    sql_client: SqlClient,
+) -> None:
+    query_spec = MetricFlowQuerySpec(
+        time_dimension_specs=(
+            MTD_SPEC_DAY,
+            metric_time_with_custom_grain,
+            normal_time_dim_with_custom_grain2,
+            TimeDimensionSpec(
+                element_name="bio_added_ts",
+                time_granularity=ExpandedTimeGranularity.from_time_granularity(TimeGranularity.MONTH),
+                entity_links=(EntityReference("user"),),
+            ),
+        ),
+    )
+
+    render_and_check(
+        request=request,
+        mf_test_configuration=mf_test_configuration,
+        dataflow_to_sql_converter=dataflow_to_sql_converter,
+        sql_client=sql_client,
+        dataflow_plan_builder=dataflow_plan_builder,
+        query_spec=query_spec,
+    )
+
+
+# TODO: Fix bug in this scenario. In optimized SQL, subq_3 alias is used but never defined.
+@pytest.mark.sql_engine_snapshot
+def test_no_metric_custom_granularity_metric_time(  # noqa: D103
+    request: FixtureRequest,
+    mf_test_configuration: MetricFlowTestConfiguration,
+    dataflow_plan_builder: DataflowPlanBuilder,
+    dataflow_to_sql_converter: DataflowToSqlQueryPlanConverter,
+    sql_client: SqlClient,
+) -> None:
+    query_spec = MetricFlowQuerySpec(
+        time_dimension_specs=(metric_time_with_custom_grain,),
+    )
+
+    render_and_check(
+        request=request,
+        mf_test_configuration=mf_test_configuration,
+        dataflow_to_sql_converter=dataflow_to_sql_converter,
+        sql_client=sql_client,
+        dataflow_plan_builder=dataflow_plan_builder,
+        query_spec=query_spec,
+    )

tests_metricflow/snapshots/test_custom_granularity.py/SqlQueryPlan/BigQuery/test_no_metric_custom_granularity_joined_to_non_default_grain__plan0_optimized.sql

@@ -0,0 +1,25 @@
+-- Join Standard Outputs
+-- Join to Custom Granularity Dataset
+-- Join to Custom Granularity Dataset
+-- Pass Only Elements: ['metric_time__day', 'metric_time__martian_day', 'user__bio_added_ts__martian_day', 'user__bio_added_ts__month']
+SELECT
+  subq_12.martian_day AS user__bio_added_ts__martian_day
+  , subq_11.martian_day AS metric_time__martian_day
+  , DATETIME_TRUNC(users_ds_source_src_28000.bio_added_ts, month) AS user__bio_added_ts__month
+  , DATETIME_TRUNC(time_spine_src_28006.ds, day) AS metric_time__day
+FROM ***************************.dim_users users_ds_source_src_28000
+CROSS JOIN
+  ***************************.mf_time_spine time_spine_src_28006
+LEFT OUTER JOIN
+  ***************************.mf_time_spine subq_11
+ON
+  subq_7.metric_time__day = subq_11.ds
+LEFT OUTER JOIN
+  ***************************.mf_time_spine subq_12
+ON
+  DATETIME_TRUNC(users_ds_source_src_28000.bio_added_ts, day) = subq_12.ds
+GROUP BY
+  user__bio_added_ts__martian_day
+  , metric_time__martian_day
+  , user__bio_added_ts__month
+  , metric_time__day

tests_metricflow/snapshots/test_custom_granularity.py/SqlQueryPlan/BigQuery/test_no_metric_custom_granularity_metric_time__plan0.sql

@@ -0,0 +1,56 @@
+-- Pass Only Elements: ['metric_time__martian_day',]
+SELECT
+  subq_2.metric_time__martian_day
+FROM (
+  -- Metric Time Dimension 'ds'
+  -- Join to Custom Granularity Dataset
+  SELECT
+    subq_0.ds__day AS ds__day
+    , subq_0.ds__week AS ds__week
+    , subq_0.ds__month AS ds__month
+    , subq_0.ds__quarter AS ds__quarter
+    , subq_0.ds__year AS ds__year
+    , subq_0.ds__extract_year AS ds__extract_year
+    , subq_0.ds__extract_quarter AS ds__extract_quarter
+    , subq_0.ds__extract_month AS ds__extract_month
+    , subq_0.ds__extract_day AS ds__extract_day
+    , subq_0.ds__extract_dow AS ds__extract_dow
+    , subq_0.ds__extract_doy AS ds__extract_doy
+    , subq_0.ds__martian_day AS ds__martian_day
+    , subq_0.ds__day AS metric_time__day
+    , subq_0.ds__week AS metric_time__week
+    , subq_0.ds__month AS metric_time__month
+    , subq_0.ds__quarter AS metric_time__quarter
+    , subq_0.ds__year AS metric_time__year
+    , subq_0.ds__extract_year AS metric_time__extract_year
+    , subq_0.ds__extract_quarter AS metric_time__extract_quarter
+    , subq_0.ds__extract_month AS metric_time__extract_month
+    , subq_0.ds__extract_day AS metric_time__extract_day
+    , subq_0.ds__extract_dow AS metric_time__extract_dow
+    , subq_0.ds__extract_doy AS metric_time__extract_doy
+    , subq_0.ds__martian_day AS metric_time__martian_day
+    , subq_1.martian_day AS metric_time__martian_day
+  FROM (
+    -- Time Spine
+    SELECT
+      DATETIME_TRUNC(time_spine_src_28006.ds, day) AS ds__day
+      , DATETIME_TRUNC(time_spine_src_28006.ds, isoweek) AS ds__week
+      , DATETIME_TRUNC(time_spine_src_28006.ds, month) AS ds__month
+      , DATETIME_TRUNC(time_spine_src_28006.ds, quarter) AS ds__quarter
+      , DATETIME_TRUNC(time_spine_src_28006.ds, year) AS ds__year
+      , EXTRACT(year FROM time_spine_src_28006.ds) AS ds__extract_year
+      , EXTRACT(quarter FROM time_spine_src_28006.ds) AS ds__extract_quarter
+      , EXTRACT(month FROM time_spine_src_28006.ds) AS ds__extract_month
+      , EXTRACT(day FROM time_spine_src_28006.ds) AS ds__extract_day
+      , IF(EXTRACT(dayofweek FROM time_spine_src_28006.ds) = 1, 7, EXTRACT(dayofweek FROM time_spine_src_28006.ds) - 1) AS ds__extract_dow
+      , EXTRACT(dayofyear FROM time_spine_src_28006.ds) AS ds__extract_doy
+      , time_spine_src_28006.martian_day AS ds__martian_day
+    FROM ***************************.mf_time_spine time_spine_src_28006
+  ) subq_0
+  LEFT OUTER JOIN
+    ***************************.mf_time_spine subq_1
+  ON
+    subq_0.metric_time__day = subq_1.ds
+) subq_2
+GROUP BY
+  metric_time__martian_day

tests_metricflow/snapshots/test_custom_granularity.py/SqlQueryPlan/BigQuery/test_no_metric_custom_granularity_metric_time__plan0_optimized.sql

@@ -0,0 +1,12 @@
+-- Metric Time Dimension 'ds'
+-- Join to Custom Granularity Dataset
+-- Pass Only Elements: ['metric_time__martian_day',]
+SELECT
+  subq_4.martian_day AS metric_time__martian_day
+FROM ***************************.mf_time_spine time_spine_src_28006
+LEFT OUTER JOIN
+  ***************************.mf_time_spine subq_4
+ON
+  subq_3.metric_time__day = subq_4.ds
+GROUP BY
+  metric_time__martian_day

tests_metricflow/snapshots/test_custom_granularity.py/SqlQueryPlan/BigQuery/test_no_metric_custom_granularity_non_metric_time__plan0.sql

@@ -0,0 +1,104 @@
+-- Pass Only Elements: ['booking__ds__martian_day',]
+SELECT
+  subq_1.booking__ds__martian_day
+FROM (
+  -- Read Elements From Semantic Model 'bookings_source'
+  -- Join to Custom Granularity Dataset
+  SELECT
+    1 AS bookings
+    , CASE WHEN is_instant THEN 1 ELSE 0 END AS instant_bookings
+    , bookings_source_src_28000.booking_value AS booking_value
+    , bookings_source_src_28000.booking_value AS max_booking_value
+    , bookings_source_src_28000.booking_value AS min_booking_value
+    , bookings_source_src_28000.guest_id AS bookers
+    , bookings_source_src_28000.booking_value AS average_booking_value
+    , bookings_source_src_28000.booking_value AS booking_payments
+    , CASE WHEN referrer_id IS NOT NULL THEN 1 ELSE 0 END AS referred_bookings
+    , bookings_source_src_28000.booking_value AS median_booking_value
+    , bookings_source_src_28000.booking_value AS booking_value_p99
+    , bookings_source_src_28000.booking_value AS discrete_booking_value_p99
+    , bookings_source_src_28000.booking_value AS approximate_continuous_booking_value_p99
+    , bookings_source_src_28000.booking_value AS approximate_discrete_booking_value_p99
+    , bookings_source_src_28000.is_instant AS is_instant
+    , DATETIME_TRUNC(bookings_source_src_28000.ds, day) AS ds__day
+    , DATETIME_TRUNC(bookings_source_src_28000.ds, isoweek) AS ds__week
+    , DATETIME_TRUNC(bookings_source_src_28000.ds, month) AS ds__month
+    , DATETIME_TRUNC(bookings_source_src_28000.ds, quarter) AS ds__quarter
+    , DATETIME_TRUNC(bookings_source_src_28000.ds, year) AS ds__year
+    , EXTRACT(year FROM bookings_source_src_28000.ds) AS ds__extract_year
+    , EXTRACT(quarter FROM bookings_source_src_28000.ds) AS ds__extract_quarter
+    , EXTRACT(month FROM bookings_source_src_28000.ds) AS ds__extract_month
+    , EXTRACT(day FROM bookings_source_src_28000.ds) AS ds__extract_day
+    , IF(EXTRACT(dayofweek FROM bookings_source_src_28000.ds) = 1, 7, EXTRACT(dayofweek FROM bookings_source_src_28000.ds) - 1) AS ds__extract_dow
+    , EXTRACT(dayofyear FROM bookings_source_src_28000.ds) AS ds__extract_doy
+    , DATETIME_TRUNC(bookings_source_src_28000.ds_partitioned, day) AS ds_partitioned__day
+    , DATETIME_TRUNC(bookings_source_src_28000.ds_partitioned, isoweek) AS ds_partitioned__week
+    , DATETIME_TRUNC(bookings_source_src_28000.ds_partitioned, month) AS ds_partitioned__month
+    , DATETIME_TRUNC(bookings_source_src_28000.ds_partitioned, quarter) AS ds_partitioned__quarter
+    , DATETIME_TRUNC(bookings_source_src_28000.ds_partitioned, year) AS ds_partitioned__year
+    , EXTRACT(year FROM bookings_source_src_28000.ds_partitioned) AS ds_partitioned__extract_year
+    , EXTRACT(quarter FROM bookings_source_src_28000.ds_partitioned) AS ds_partitioned__extract_quarter
+    , EXTRACT(month FROM bookings_source_src_28000.ds_partitioned) AS ds_partitioned__extract_month
+    , EXTRACT(day FROM bookings_source_src_28000.ds_partitioned) AS ds_partitioned__extract_day
+    , IF(EXTRACT(dayofweek FROM bookings_source_src_28000.ds_partitioned) = 1, 7, EXTRACT(dayofweek FROM bookings_source_src_28000.ds_partitioned) - 1) AS ds_partitioned__extract_dow
+    , EXTRACT(dayofyear FROM bookings_source_src_28000.ds_partitioned) AS ds_partitioned__extract_doy
+    , DATETIME_TRUNC(bookings_source_src_28000.paid_at, day) AS paid_at__day
+    , DATETIME_TRUNC(bookings_source_src_28000.paid_at, isoweek) AS paid_at__week
+    , DATETIME_TRUNC(bookings_source_src_28000.paid_at, month) AS paid_at__month
+    , DATETIME_TRUNC(bookings_source_src_28000.paid_at, quarter) AS paid_at__quarter
+    , DATETIME_TRUNC(bookings_source_src_28000.paid_at, year) AS paid_at__year
+    , EXTRACT(year FROM bookings_source_src_28000.paid_at) AS paid_at__extract_year
+    , EXTRACT(quarter FROM bookings_source_src_28000.paid_at) AS paid_at__extract_quarter
+    , EXTRACT(month FROM bookings_source_src_28000.paid_at) AS paid_at__extract_month
+    , EXTRACT(day FROM bookings_source_src_28000.paid_at) AS paid_at__extract_day
+    , IF(EXTRACT(dayofweek FROM bookings_source_src_28000.paid_at) = 1, 7, EXTRACT(dayofweek FROM bookings_source_src_28000.paid_at) - 1) AS paid_at__extract_dow
+    , EXTRACT(dayofyear FROM bookings_source_src_28000.paid_at) AS paid_at__extract_doy
+    , bookings_source_src_28000.is_instant AS booking__is_instant
+    , DATETIME_TRUNC(bookings_source_src_28000.ds, day) AS booking__ds__day
+    , DATETIME_TRUNC(bookings_source_src_28000.ds, isoweek) AS booking__ds__week
+    , DATETIME_TRUNC(bookings_source_src_28000.ds, month) AS booking__ds__month
+    , DATETIME_TRUNC(bookings_source_src_28000.ds, quarter) AS booking__ds__quarter
+    , DATETIME_TRUNC(bookings_source_src_28000.ds, year) AS booking__ds__year
+    , EXTRACT(year FROM bookings_source_src_28000.ds) AS booking__ds__extract_year
+    , EXTRACT(quarter FROM bookings_source_src_28000.ds) AS booking__ds__extract_quarter
+    , EXTRACT(month FROM bookings_source_src_28000.ds) AS booking__ds__extract_month
+    , EXTRACT(day FROM bookings_source_src_28000.ds) AS booking__ds__extract_day
+    , IF(EXTRACT(dayofweek FROM bookings_source_src_28000.ds) = 1, 7, EXTRACT(dayofweek FROM bookings_source_src_28000.ds) - 1) AS booking__ds__extract_dow
+    , EXTRACT(dayofyear FROM bookings_source_src_28000.ds) AS booking__ds__extract_doy
+    , DATETIME_TRUNC(bookings_source_src_28000.ds_partitioned, day) AS booking__ds_partitioned__day
+    , DATETIME_TRUNC(bookings_source_src_28000.ds_partitioned, isoweek) AS booking__ds_partitioned__week
+    , DATETIME_TRUNC(bookings_source_src_28000.ds_partitioned, month) AS booking__ds_partitioned__month
+    , DATETIME_TRUNC(bookings_source_src_28000.ds_partitioned, quarter) AS booking__ds_partitioned__quarter
+    , DATETIME_TRUNC(bookings_source_src_28000.ds_partitioned, year) AS booking__ds_partitioned__year
+    , EXTRACT(year FROM bookings_source_src_28000.ds_partitioned) AS booking__ds_partitioned__extract_year
+    , EXTRACT(quarter FROM bookings_source_src_28000.ds_partitioned) AS booking__ds_partitioned__extract_quarter
+    , EXTRACT(month FROM bookings_source_src_28000.ds_partitioned) AS booking__ds_partitioned__extract_month
+    , EXTRACT(day FROM bookings_source_src_28000.ds_partitioned) AS booking__ds_partitioned__extract_day
+    , IF(EXTRACT(dayofweek FROM bookings_source_src_28000.ds_partitioned) = 1, 7, EXTRACT(dayofweek FROM bookings_source_src_28000.ds_partitioned) - 1) AS booking__ds_partitioned__extract_dow
+    , EXTRACT(dayofyear FROM bookings_source_src_28000.ds_partitioned) AS booking__ds_partitioned__extract_doy
+    , DATETIME_TRUNC(bookings_source_src_28000.paid_at, day) AS booking__paid_at__day
+    , DATETIME_TRUNC(bookings_source_src_28000.paid_at, isoweek) AS booking__paid_at__week
+    , DATETIME_TRUNC(bookings_source_src_28000.paid_at, month) AS booking__paid_at__month
+    , DATETIME_TRUNC(bookings_source_src_28000.paid_at, quarter) AS booking__paid_at__quarter
+    , DATETIME_TRUNC(bookings_source_src_28000.paid_at, year) AS booking__paid_at__year
+    , EXTRACT(year FROM bookings_source_src_28000.paid_at) AS booking__paid_at__extract_year
+    , EXTRACT(quarter FROM bookings_source_src_28000.paid_at) AS booking__paid_at__extract_quarter
+    , EXTRACT(month FROM bookings_source_src_28000.paid_at) AS booking__paid_at__extract_month
+    , EXTRACT(day FROM bookings_source_src_28000.paid_at) AS booking__paid_at__extract_day
+    , IF(EXTRACT(dayofweek FROM bookings_source_src_28000.paid_at) = 1, 7, EXTRACT(dayofweek FROM bookings_source_src_28000.paid_at) - 1) AS booking__paid_at__extract_dow
+    , EXTRACT(dayofyear FROM bookings_source_src_28000.paid_at) AS booking__paid_at__extract_doy
+    , bookings_source_src_28000.listing_id AS listing
+    , bookings_source_src_28000.guest_id AS guest
+    , bookings_source_src_28000.host_id AS host
+    , bookings_source_src_28000.listing_id AS booking__listing
+    , bookings_source_src_28000.guest_id AS booking__guest
+    , bookings_source_src_28000.host_id AS booking__host
+    , subq_0.martian_day AS booking__ds__martian_day
+  FROM ***************************.fct_bookings bookings_source_src_28000
+  LEFT OUTER JOIN
+    ***************************.mf_time_spine subq_0
+  ON
+    bookings_source_src_28000.booking__ds__day = subq_0.ds
+) subq_1
+GROUP BY
+  booking__ds__martian_day

tests_metricflow/snapshots/test_custom_granularity.py/SqlQueryPlan/BigQuery/test_no_metric_custom_granularity_non_metric_time__plan0_optimized.sql

@@ -0,0 +1,12 @@
+-- Read Elements From Semantic Model 'bookings_source'
+-- Join to Custom Granularity Dataset
+-- Pass Only Elements: ['booking__ds__martian_day',]
+SELECT
+  subq_2.martian_day AS booking__ds__martian_day
+FROM ***************************.fct_bookings bookings_source_src_28000
+LEFT OUTER JOIN
+  ***************************.mf_time_spine subq_2
+ON
+  bookings_source_src_28000.booking__ds__day = subq_2.ds
+GROUP BY
+  booking__ds__martian_day