added tests for constant properties

dbt-labs · Jan 19, 2023 · 348b4f3 · 348b4f3
1 parent ac82e5c
commit 348b4f3
Show file tree

Hide file tree

Showing 5 changed files with 97 additions and 29 deletions.
diff --git a/metricflow/test/fixtures/model_yamls/simple_model/data_sources/buys_source.yaml b/metricflow/test/fixtures/model_yamls/simple_model/data_sources/buys_source.yaml
@@ -28,3 +28,5 @@ data_source:
     - name: user
       type: foreign
       expr: user_id
+    - name: session_id
+      type: foreign
diff --git a/metricflow/test/fixtures/model_yamls/simple_model/data_sources/visits_source.yaml b/metricflow/test/fixtures/model_yamls/simple_model/data_sources/visits_source.yaml
@@ -30,4 +30,7 @@ data_source:
   identifiers:
     - name: user
       type: foreign
-      expr: user_id
+      expr: user_id
+    - name: session
+      type: foreign
+      expr: session_id
diff --git a/metricflow/test/fixtures/model_yamls/simple_model/metrics.yaml b/metricflow/test/fixtures/model_yamls/simple_model/metrics.yaml
@@ -710,3 +710,20 @@ metric:
       window: 7 days
       entity: user
       calculation: conversions
+---
+metric:
+  name: visit_buy_conversion_rate_by_session
+  description: conversion rate on visits-buys on a 7 day window held by a constant session_id
+  owners:
+    - [email protected]
+  type: conversion
+  type_params:
+    conversion_type_params:
+      base_measure: visits
+      conversion_measure: buys
+      window: 7 days
+      entity: user
+      calculation: conversion_rate
+      constant_properties:
+        - name: session_id
+          base_expr: session
diff --git a/metricflow/test/fixtures/table_fixtures.py b/metricflow/test/fixtures/table_fixtures.py
@@ -331,54 +331,54 @@ def create_simple_model_tables(mf_test_session_state: MetricFlowTestSessionState
 
     # Events data
     visits_data = [
-        ("u0004114", "2020-01-01", "fb_ad_1"),
-        ("u0004214", "2020-01-01", "fb_ad_2"),
-        ("u0003141", "2020-01-01", "homepage_1"),
-        ("u0003154", "2020-01-01", "homepage_2"),
-        ("u1612112", "2020-01-02", "homepage_1"),
-        ("u0042324", "2020-01-02", "homepage_1"),
-        ("u0005432", "2020-01-02", "fb_ad_3"),
-        ("u0003452", "2020-01-02", "fb_ad_1"),
-        ("u0003452", "2020-01-02", "user_2"),
-        ("u0042324", "2020-01-03", "fb_ad_1"),
-        ("u0005432", "2020-01-03", "google_ad_1"),
-        ("u0005472", "2020-01-03", "google_ad_2"),
-        ("u0005414", "2020-01-04", "google_ad_1"),
-        ("u0004114", "2020-01-06", "homepage_1"),
-        ("u0004114", "2020-01-07", "fb_ad_2"),
-        ("u0004117", "2020-01-10", "google_ad_1"),
-        ("u0003141", "2020-01-12", "user_1"),
+        ("u0004114", "2020-01-01", "fb_ad_1", "s1"),
+        ("u0004214", "2020-01-01", "fb_ad_2", "s2"),
+        ("u0003141", "2020-01-01", "homepage_1", "s3"),
+        ("u0003154", "2020-01-01", "homepage_2", "s4"),
+        ("u1612112", "2020-01-02", "homepage_1", "s5"),
+        ("u0042324", "2020-01-02", "homepage_1", "s6"),
+        ("u0005432", "2020-01-02", "fb_ad_3", "s7"),
+        ("u0003452", "2020-01-02", "fb_ad_1", "s8"),
+        ("u0003452", "2020-01-02", "user_2", "s8"),
+        ("u0042324", "2020-01-03", "fb_ad_1", "s9"),
+        ("u0005432", "2020-01-03", "google_ad_1", "s10"),
+        ("u0005472", "2020-01-03", "google_ad_2", "s11"),
+        ("u0005414", "2020-01-04", "google_ad_1", "s12"),
+        ("u0004114", "2020-01-06", "homepage_1", "s13"),
+        ("u0004114", "2020-01-07", "fb_ad_2", "s14"),
+        ("u0004117", "2020-01-10", "google_ad_1", "s15"),
+        ("u0003141", "2020-01-12", "user_1", "s16"),
     ]
 
     create_table(
         sql_client=sql_client,
         sql_table=SqlTable(schema_name=schema, table_name="fct_visits"),
         df=make_df(
             sql_client=sql_client,
-            columns=["user_id", DEFAULT_DS, "referrer_id"],
+            columns=["user_id", DEFAULT_DS, "referrer_id", "session_id"],
             time_columns={DEFAULT_DS},
             data=visits_data,
         ),
     )
 
     buy_data = [
-        ("u0004114", "2020-01-02"),
-        ("u0042324", "2020-01-03"),
-        ("u0042324", "2020-01-03"),
-        ("u0005432", "2020-01-04"),
-        ("u1612112", "2020-01-07"),
-        ("u0004114", "2020-01-07"),
-        ("u0004117", "2020-01-10"),
-        ("u0003141", "2020-01-07"),
-        ("u0003452", "2020-01-04"),
+        ("u0004114", "2020-01-02", "s1"),
+        ("u0042324", "2020-01-03", "s6"),
+        ("u0042324", "2020-01-03", "s9"),
+        ("u0005432", "2020-01-04", "s7"),
+        ("u1612112", "2020-01-07", "s0"),
+        ("u0004114", "2020-01-07", "s14"),
+        ("u0004117", "2020-01-10", "s15"),
+        ("u0003141", "2020-01-07", "s3"),
+        ("u0003452", "2020-01-04", "s8"),
     ]
 
     create_table(
         sql_client=sql_client,
         sql_table=SqlTable(schema_name=schema, table_name="fct_buys"),
         df=make_df(
             sql_client=sql_client,
-            columns=["user_id", DEFAULT_DS],
+            columns=["user_id", DEFAULT_DS, "session_id"],
             time_columns={DEFAULT_DS},
             data=buy_data,
         ),

diff --git a/metricflow/test/integration/test_cases/itest_conversion_metrics.yaml b/metricflow/test/integration/test_cases/itest_conversion_metrics.yaml
@@ -228,3 +228,49 @@ integration_test:
       GROUP BY
         b.referrer_id, b.ds
     ) conversions
+---
+integration_test:
+  name: conversion_rate_metric_with_constant_property
+  description: Query a conversion metric that calculates the conversion rate held by a constant property
+  model: SIMPLE_MODEL
+  metrics: ["visit_buy_conversion_rate_by_session"]
+  group_bys: ["metric_time"]
+  check_query: |
+    SELECT
+      opportunities.metric_time AS metric_time
+      , CAST(conversions.buys AS {{ double_data_type_name }}) / CAST(NULLIF(opportunities.visits, 0) AS {{ double_data_type_name }}) AS visit_buy_conversion_rate_by_session
+    FROM (
+      SELECT
+        metric_time, SUM(a.visits) AS visits
+      FROM (
+        SELECT
+          ds AS metric_time, 1 AS visits
+        FROM {{ source_schema }}.fct_visits visits
+      ) a
+      GROUP BY
+        a.metric_time
+    ) opportunities
+    INNER JOIN (
+      SELECT
+        b.ds AS metric_time, SUM(b.buys) AS buys
+      FROM (
+        SELECT DISTINCT
+        first_value(v.ds) OVER (PARTITION BY buy_source.ds, buy_source.user_id, buy_source.session_id ORDER BY v.ds DESC NULLS FIRST) AS ds
+        , first_value(v.user_id) OVER (PARTITION BY buy_source.ds, buy_source.user_id, buy_source.session_id ORDER BY v.ds DESC NULLS FIRST) AS user_id
+        , first_value(v.referrer_id) OVER (PARTITION BY buy_source.ds, buy_source.user_id, buy_source.session_id ORDER BY v.ds DESC NULLS FIRST) AS referrer_id
+        , buy_source.uuid
+        , 1 AS buys
+        FROM {{ source_schema }}.fct_visits v
+        INNER JOIN
+        (
+          SELECT *, {{ generate_random_uuid() }} AS uuid FROM {{ source_schema }}.fct_buys
+        ) buy_source
+        ON
+          v.user_id = buy_source.user_id 
+          AND v.ds <= buy_source.ds AND v.ds > {{ render_date_sub("buy_source", "ds", 7, TimeGranularity.DAY) }}
+          AND buy_source.session_id = v.session_id
+      ) b
+      GROUP BY
+        b.ds
+    ) conversions
+    ON opportunities.metric_time = conversions.metric_time