From ff841c32fa9bc7849c7d38506ca5cd56f75321dc Mon Sep 17 00:00:00 2001
From: Claire Carroll <claire@fishtownanalytics.com>
Date: Wed, 10 Mar 2021 10:52:02 -0500
Subject: [PATCH] Add fewer_rows_than schema test (#343)

---
 CHANGELOG.md                                  |  1 +
 README.md                                     | 14 ++++++
 .../data_test_fewer_rows_than_table_1.csv     |  4 ++
 .../data_test_fewer_rows_than_table_2.csv     |  5 +++
 .../schema_tests/test_fewer_rows_than.sql     |  9 ++++
 macros/schema_tests/fewer_rows_than.sql       | 43 +++++++++++++++++++
 6 files changed, 76 insertions(+)
 create mode 100644 integration_tests/data/schema_tests/data_test_fewer_rows_than_table_1.csv
 create mode 100644 integration_tests/data/schema_tests/data_test_fewer_rows_than_table_2.csv
 create mode 100644 integration_tests/models/schema_tests/test_fewer_rows_than.sql
 create mode 100644 macros/schema_tests/fewer_rows_than.sql

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 46e03eb2..9bc34c80 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -7,6 +7,7 @@
 * Add new schema test, `sequential_values` ([#318](https://github.com/fishtown-analytics/dbt-utils/pull/318), inspired by [@hundredwatt](https://github.com/hundredwatt))
 * Support `quarter` in the `postgres__last_day` macro ([#333](https://github.com/fishtown-analytics/dbt-utils/pull/333/files), [@seunghanhong](https://github.com/seunghanhong))
 * Add new argument, `unit`, to `haversine_distance` [#340](https://github.com/fishtown-analytics/dbt-utils/pull/340) [@bastienboutonnet](https://github.com/bastienboutonnet)
+* Add new schema test, `fewer_rows_than` (code originally in [#221](https://github.com/fishtown-analytics/dbt-utils/pull/230/) from [@dmarts](https://github.com/dmarts), merged via [#343])
 
 
 ## Fixes
diff --git a/README.md b/README.md
index 61a853a1..61a093d0 100644
--- a/README.md
+++ b/README.md
@@ -138,6 +138,20 @@ models:
 
 ```
 
+#### fewer_rows_than ([source](macros/schema_tests/fewer_rows_than.sql))
+This schema test asserts that this model has fewer rows than the referenced model.
+
+Usage:
+```yaml
+version: 2
+
+models:
+  - name: model_name
+    tests:
+      - dbt_utils.fewer_rows_than:
+          compare_model: ref('other_table_name')
+```
+
 #### equality ([source](macros/schema_tests/equality.sql))
 This schema test asserts the equality of two relations. Optionally specify a subset of columns to compare.
 
diff --git a/integration_tests/data/schema_tests/data_test_fewer_rows_than_table_1.csv b/integration_tests/data/schema_tests/data_test_fewer_rows_than_table_1.csv
new file mode 100644
index 00000000..ba0e5ec6
--- /dev/null
+++ b/integration_tests/data/schema_tests/data_test_fewer_rows_than_table_1.csv
@@ -0,0 +1,4 @@
+field
+1
+2
+3
diff --git a/integration_tests/data/schema_tests/data_test_fewer_rows_than_table_2.csv b/integration_tests/data/schema_tests/data_test_fewer_rows_than_table_2.csv
new file mode 100644
index 00000000..eb0f035a
--- /dev/null
+++ b/integration_tests/data/schema_tests/data_test_fewer_rows_than_table_2.csv
@@ -0,0 +1,5 @@
+field
+1
+2
+3
+4
diff --git a/integration_tests/models/schema_tests/test_fewer_rows_than.sql b/integration_tests/models/schema_tests/test_fewer_rows_than.sql
new file mode 100644
index 00000000..c2ad4cbd
--- /dev/null
+++ b/integration_tests/models/schema_tests/test_fewer_rows_than.sql
@@ -0,0 +1,9 @@
+with data as (
+
+    select * from {{ ref('data_test_fewer_rows_than_table_1') }}
+
+)
+
+select
+    field
+from data
\ No newline at end of file
diff --git a/macros/schema_tests/fewer_rows_than.sql b/macros/schema_tests/fewer_rows_than.sql
new file mode 100644
index 00000000..328c0ea0
--- /dev/null
+++ b/macros/schema_tests/fewer_rows_than.sql
@@ -0,0 +1,43 @@
+{% macro test_fewer_rows_than(model) %}
+  {{ return(adapter.dispatch('test_fewer_rows_than', packages = dbt_utils._get_utils_namespaces())(model, combination_of_columns, quote_columns, where)) }}
+{% endmacro %}
+
+{% macro default__test_fewer_rows_than(model) %}
+
+{% set compare_model = kwargs.get('compare_model', kwargs.get('arg')) %}
+
+with a as (
+
+    select count(*) as count_ourmodel from {{ model }}
+
+),
+b as (
+
+    select count(*) as count_comparisonmodel from {{ compare_model }}
+
+),
+counts as (
+
+    select
+        (select count_ourmodel from a) as count_model_with_fewer_rows,
+        (select count_comparisonmodel from b) as count_model_with_more_rows
+
+),
+final as (
+
+    select
+        case
+            -- fail the test if we have more rows than the reference model and return the row count delta
+            when count_model_with_fewer_rows > count_model_with_more_rows then (count_model_with_fewer_rows - count_model_with_more_rows)
+            -- fail the test if they are the same number
+            when count_model = count_comparison then 1
+            -- pass the test if the delta is positive (i.e. return the number 0)
+            else 0
+    end as row_count_delta
+    from counts
+
+)
+
+select row_count_delta from final
+
+{% endmacro %}