From 96c461b65822440f173ab1a68f1302f96e99aed2 Mon Sep 17 00:00:00 2001
From: Judah Rand <17158624+judahrand@users.noreply.github.com>
Date: Fri, 25 Mar 2022 03:52:16 +0000
Subject: [PATCH] Add `alias` argument to `deduplicate` macro (#526)

* Add `alias` argument to `deduplicate

* Test `alias` argument

* Rename `alias` to `relation_alias`
---
 README.md                                     |  5 +++--
 .../data/sql/data_deduplicate.csv             |  1 +
 .../models/sql/test_deduplicate.sql           | 19 +++++++++++++++++--
 macros/sql/deduplicate.sql                    | 12 ++++++------
 4 files changed, 27 insertions(+), 10 deletions(-)

diff --git a/README.md b/README.md
index acc7c825..80c6cbbb 100644
--- a/README.md
+++ b/README.md
@@ -36,7 +36,7 @@ For compatibility details between versions of dbt-core and dbt-utils, [see this
 
 - [SQL generators](#sql-generators)
     - [date_spine](#date_spine-source)
-    - [dedupe](#dedupe-source)
+    - [deduplicate](#deduplicate)
     - [haversine_distance](#haversine_distance-source)
     - [group_by](#group_by-source)
     - [star](#star-source)
@@ -715,7 +715,8 @@ This macro returns the sql required to remove duplicate rows from a model or sou
 {{ dbt_utils.deduplicate(
     relation=source('my_source', 'my_table'),
     group_by="user_id, cast(timestamp as day)",
-    order_by="timestamp desc"
+    order_by="timestamp desc",
+    relation_alias="my_cte"
    )
 }}
 ```
diff --git a/integration_tests/data/sql/data_deduplicate.csv b/integration_tests/data/sql/data_deduplicate.csv
index c3ae0c4d..7e06170a 100644
--- a/integration_tests/data/sql/data_deduplicate.csv
+++ b/integration_tests/data/sql/data_deduplicate.csv
@@ -1,3 +1,4 @@
 user_id,event,version
 1,play,1
 1,play,2
+2,pause,1
diff --git a/integration_tests/models/sql/test_deduplicate.sql b/integration_tests/models/sql/test_deduplicate.sql
index 7df79261..81fe81e7 100644
--- a/integration_tests/models/sql/test_deduplicate.sql
+++ b/integration_tests/models/sql/test_deduplicate.sql
@@ -1,6 +1,21 @@
-with deduped as (
+with
 
-    {{ dbt_utils.deduplicate(ref('data_deduplicate'), group_by='user_id', order_by='version desc') | indent }}
+source as (
+    select *
+    from {{ ref('data_deduplicate') }}
+    where user_id = 1
+),
+
+deduped as (
+
+    {{
+        dbt_utils.deduplicate(
+            ref('data_deduplicate'),
+            group_by='user_id',
+            order_by='version desc',
+            relation_alias="source"
+        ) | indent
+    }}
 
 )
 
diff --git a/macros/sql/deduplicate.sql b/macros/sql/deduplicate.sql
index f5d65534..9a3571a2 100644
--- a/macros/sql/deduplicate.sql
+++ b/macros/sql/deduplicate.sql
@@ -1,8 +1,8 @@
-{%- macro deduplicate(relation, group_by, order_by=none) -%}
-    {{ return(adapter.dispatch('deduplicate', 'dbt_utils')(relation, group_by, order_by=order_by)) }}
+{%- macro deduplicate(relation, group_by, order_by=none, relation_alias=none) -%}
+    {{ return(adapter.dispatch('deduplicate', 'dbt_utils')(relation, group_by, order_by=order_by, relation_alias=relation_alias)) }}
 {% endmacro %}
 
-{%- macro default__deduplicate(relation, group_by, order_by=none) -%}
+{%- macro default__deduplicate(relation, group_by, order_by=none, relation_alias=none) -%}
 
     select
         {{ dbt_utils.star(relation, relation_alias='deduped') | indent }}
@@ -15,7 +15,7 @@
                 order by {{ order_by }}
                 {%- endif %}
             ) as rn
-        from {{ relation }} as _inner
+        from {{ relation if relation_alias is none else relation_alias }} as _inner
     ) as deduped
     where deduped.rn = 1
 
@@ -26,7 +26,7 @@
 --  clause in BigQuery:
 --  https://github.com/dbt-labs/dbt-utils/issues/335#issuecomment-788157572
 #}
-{%- macro bigquery__deduplicate(relation, group_by, order_by=none) -%}
+{%- macro bigquery__deduplicate(relation, group_by, order_by=none, relation_alias=none) -%}
 
     select
         {{ dbt_utils.star(relation, relation_alias='deduped') | indent }}
@@ -39,7 +39,7 @@
                 {%- endif %}
                 limit 1
             )[offset(0)] as deduped
-        from {{ relation }} as original
+        from {{ relation if relation_alias is none else relation_alias }} as original
         group by {{ group_by }}
     )