From 873381f145af08922fda049c1928e33fd9ecb930 Mon Sep 17 00:00:00 2001
From: Manrique Vargas <mv1742@nyu.edu>
Date: Fri, 13 Oct 2023 20:16:16 -0600
Subject: [PATCH] add seed (#3)

* insert overwrite instead of insert into for new seed runs

* Mv1742 load csv table seed (#1)

* changie commit

* Update Fixes-20231013-120628.yaml

* insert overwrite instead of insert into for new seed runs

changie commit

Mv1742 load csv table seed (#1)

* changie commit

* Update Fixes-20231013-120628.yaml

* Mv1742 rebase truncate (#2)

* merge remote

* add docs-issue workflow to dbt-spark (#913)

* Update seed.sql

* Update seed.sql

* add truncate table function

* rm changelog

* merge remote
---
 .changes/unreleased/Fixes-20231013-120628.yaml             | 7 +++++++
 dbt/include/spark/macros/adapters.sql                      | 7 +++++++
 .../macros/materializations/incremental/incremental.sql    | 4 ++++
 dbt/include/spark/macros/materializations/seed.sql         | 5 ++++-
 4 files changed, 22 insertions(+), 1 deletion(-)
 create mode 100644 .changes/unreleased/Fixes-20231013-120628.yaml

diff --git a/.changes/unreleased/Fixes-20231013-120628.yaml b/.changes/unreleased/Fixes-20231013-120628.yaml
new file mode 100644
index 000000000..a448944ad
--- /dev/null
+++ b/.changes/unreleased/Fixes-20231013-120628.yaml
@@ -0,0 +1,7 @@
+kind: Fixes
+body: Overwrite existing rows on existing seed tables. For unmanaged databases (no location specified), the current seed command in
+  dbt-spark appends to existing seeded tables instead overwriting.
+time: 2023-10-13T12:06:28.078483-06:00
+custom:
+  Author: mv1742
+  Issue: "112"
diff --git a/dbt/include/spark/macros/adapters.sql b/dbt/include/spark/macros/adapters.sql
index bfc1f198d..461e8f14f 100644
--- a/dbt/include/spark/macros/adapters.sql
+++ b/dbt/include/spark/macros/adapters.sql
@@ -342,6 +342,13 @@
   {%- endcall %}
 {% endmacro %}
 
+
+{% macro spark__truncate_relation(relation) -%}
+  {% call statement('truncate_relation', auto_begin=False) -%}
+    truncate {{ relation.type }} if exists {{ relation }}
+  {%- endcall %}
+{% endmacro %}
+
 {% macro spark__drop_relation(relation) -%}
   {% call statement('drop_relation', auto_begin=False) -%}
     drop {{ relation.type }} if exists {{ relation }}
diff --git a/dbt/include/spark/macros/materializations/incremental/incremental.sql b/dbt/include/spark/macros/materializations/incremental/incremental.sql
index 10d4f3ed8..1e0531c58 100644
--- a/dbt/include/spark/macros/materializations/incremental/incremental.sql
+++ b/dbt/include/spark/macros/materializations/incremental/incremental.sql
@@ -66,6 +66,10 @@
       re: python models and temporary views.
 
       Also, why do neither drop_relation or adapter.drop_relation work here?!
+      'unmanaged' tables in spark need to manually delete the database
+      otherwise drop statement does not delete the underlying data.
+      TODO:add warning that this feature does not work for Unmanaged tables.
+      Managed tables are fine.
       --#}
       {% call statement('drop_relation') -%}
         drop table if exists {{ tmp_relation }}
diff --git a/dbt/include/spark/macros/materializations/seed.sql b/dbt/include/spark/macros/materializations/seed.sql
index 196479cb0..b2b1d6a44 100644
--- a/dbt/include/spark/macros/materializations/seed.sql
+++ b/dbt/include/spark/macros/materializations/seed.sql
@@ -5,7 +5,10 @@
 
 {% macro spark__reset_csv_table(model, full_refresh, old_relation, agate_table) %}
     {% if old_relation %}
+        {{ adapter.truncate_relation(old_relation) }}
         {{ adapter.drop_relation(old_relation) }}
+
+        {{ return(sql) }}
     {% endif %}
     {% set sql = create_csv_table(model, agate_table) %}
     {{ return(sql) }}
@@ -27,7 +30,7 @@
       {% endfor %}
 
       {% set sql %}
-          insert into {{ this.render() }} values
+          insert {% if loop.index0 == 0 -%} overwrite {% else -%} into {% endif -%} {{ this.render() }} values
           {% for row in chunk -%}
               ({%- for col_name in agate_table.column_names -%}
                   {%- set inferred_type = adapter.convert_type(agate_table, loop.index0) -%}