make full refresh a config item

Added integration tests, fixed existing tests

CHANGELOG.md

@@ -1,5 +1,8 @@
 ## dbt 0.18.0 (Release TBD)
 
+### Features
+- Added a `full_refresh` config item that overrides the behavior of the `--full-refresh` flag ([#1009](https://github.com/fishtown-analytics/dbt/issues/1009), [#2348](https://github.com/fishtown-analytics/dbt/pull/2348))
+
 ## dbt 0.17.0 (Release TBD)
 
 ## dbt 0.17.0rc1 (May 12, 2020)
@@ -28,7 +31,7 @@
 - Track distinct project hashes in anonymous usage metrics for package downloads ([#2351](https://github.com/fishtown-analytics/dbt/issues/2351), [#2429](https://github.com/fishtown-analytics/dbt/pull/2429))
 
 Contributors:
- - [@azhard](https://github.com/azhard) ([#2413](https://github.com/fishtown-analytics/dbt/pull/2413), [#2422](https://github.com/fishtown-analytics/dbt/pull/2422)) 
+ - [@azhard](https://github.com/azhard) ([#2413](https://github.com/fishtown-analytics/dbt/pull/2413), [#2422](https://github.com/fishtown-analytics/dbt/pull/2422))
  - [@mikaelene](https://github.com/mikaelene) [#2414](https://github.com/fishtown-analytics/dbt/pull/2414)
  - [@raalsky](https://github.com/Raalsky) ([#2343](https://github.com/fishtown-analytics/dbt/pull/2343))
  - [@alf-mindshift](https://github.com/alf-mindshift) ([docs#90](https://github.com/fishtown-analytics/dbt-docs/pull/90))

core/dbt/contracts/graph/model_config.py

@@ -321,6 +321,7 @@ class NodeConfig(BaseConfig):
         # TODO: hide this one?
         metadata=MergeBehavior.Append.meta(),
     )
+    full_refresh: Optional[bool] = None
 
     @classmethod
     def from_dict(cls, data, validate=True):

core/dbt/include/global_project/macros/etc/is_incremental.sql

@@ -8,6 +8,6 @@
         {{ return(relation is not none
                   and relation.type == 'table'
                   and model.config.materialized == 'incremental'
-                  and not flags.FULL_REFRESH) }}
+                  and not should_full_refresh()) }}
     {% endif %}
 {% endmacro %}

core/dbt/include/global_project/macros/materializations/helpers.sql

@@ -63,3 +63,12 @@
     identifier=relation.identifier
   )) -%}
 {% endmacro %}
+
+
+{% macro should_full_refresh() %}
+  {% set config_full_refresh = config.get('full_refresh', false) %}
+  {% if config_full_refresh is none %}
+    {% set config_full_refresh = flags.FULL_REFRESH %}
+  {% endif %}
+  {% do return(config_full_refresh) %}
+{% endmacro %}

core/dbt/include/global_project/macros/materializations/incremental/incremental.sql

@@ -2,7 +2,6 @@
 {% materialization incremental, default -%}
 
   {% set unique_key = config.get('unique_key') %}
-  {% set full_refresh_mode = flags.FULL_REFRESH %}
 
   {% set target_relation = this.incorporate(type='table') %}
   {% set existing_relation = load_relation(this) %}
@@ -16,7 +15,7 @@
   {% set to_drop = [] %}
   {% if existing_relation is none %}
       {% set build_sql = create_table_as(False, target_relation, sql) %}
-  {% elif existing_relation.is_view or full_refresh_mode %}
+  {% elif existing_relation.is_view or should_full_refresh() %}
       {#-- Make sure the backup doesn't exist so we don't encounter issues with the rename below #}
       {% set backup_identifier = existing_relation.identifier ~ "__dbt_backup" %}
       {% set backup_relation = existing_relation.incorporate(path={"identifier": backup_identifier}) %}

core/dbt/include/global_project/macros/materializations/seed/seed.sql

@@ -104,7 +104,7 @@
 {% materialization seed, default %}
 
   {%- set identifier = model['alias'] -%}
-  {%- set full_refresh_mode = (flags.FULL_REFRESH == True) -%}
+  {%- set full_refresh_mode = (should_full_refresh()) -%}
 
   {%- set old_relation = adapter.get_relation(database=database, schema=schema, identifier=identifier) -%}
 

core/dbt/include/global_project/macros/materializations/view/create_or_replace_view.sql

@@ -42,7 +42,7 @@
   -- that's an error. If we were told to full refresh, drop it. This behavior differs
   -- for Snowflake and BigQuery, so multiple dispatch is used.
   {%- if old_relation is not none and old_relation.is_table -%}
-    {{ handle_existing_table(flags.FULL_REFRESH, old_relation) }}
+    {{ handle_existing_table(should_full_refresh(), old_relation) }}
   {%- endif -%}
 
   -- build model

plugins/bigquery/dbt/include/bigquery/macros/materializations/incremental.sql

@@ -17,11 +17,11 @@
 
 {% macro bq_insert_overwrite(tmp_relation, target_relation, sql, unique_key, partition_by, partitions, dest_columns) %}
   {%- set partition_type =
-      'date' if partition_by.data_type in ('timestamp, datetime') 
+      'date' if partition_by.data_type in ('timestamp, datetime')
       else partition_by.data_type -%}
 
   {% if partitions is not none and partitions != [] %} {# static #}
-  
+
       {% set predicate -%}
           {{ partition_by.render(alias='DBT_INTERNAL_DEST') }} in (
               {{ partitions | join (', ') }}
@@ -33,11 +33,11 @@
           {{sql}}
         )
       {%- endset -%}
-      
+
       {{ get_insert_overwrite_merge_sql(target_relation, source_sql, dest_columns, [predicate], include_sql_header=true) }}
-  
+
   {% else %} {# dynamic #}
-  
+
       {% set predicate -%}
           {{ partition_by.render(alias='DBT_INTERNAL_DEST') }} in unnest(dbt_partitions_for_replacement)
       {%- endset %}
@@ -65,7 +65,7 @@
               array_agg(distinct {{ partition_by.render() }})
           from {{ tmp_relation }}
       );
-      
+
       {#
         TODO: include_sql_header is a hack; consider a better approach that includes
               the sql_header at the materialization-level instead
@@ -75,7 +75,7 @@
 
       -- 4. clean up the temp table
       drop table if exists {{ tmp_relation }}
-  
+
   {% endif %}
 
 {% endmacro %}
@@ -84,12 +84,12 @@
 {% materialization incremental, adapter='bigquery' -%}
 
   {%- set unique_key = config.get('unique_key') -%}
-  {%- set full_refresh_mode = (flags.FULL_REFRESH == True) -%}
+  {%- set full_refresh_mode = (should_full_refresh()) -%}
 
   {%- set target_relation = this %}
   {%- set existing_relation = load_relation(this) %}
   {%- set tmp_relation = make_temp_relation(this) %}
-  
+
   {#-- Validate early so we don't run SQL if the strategy is invalid --#}
   {% set strategy = dbt_bigquery_validate_get_incremental_strategy(config) -%}
 
@@ -118,14 +118,14 @@
 
      {#-- if partitioned, use BQ scripting to get the range of partition values to be updated --#}
      {% if strategy == 'insert_overwrite' %}
-     
+
         {% set missing_partition_msg -%}
           The 'insert_overwrite' strategy requires the `partition_by` config.
         {%- endset %}
         {% if partition_by is none %}
           {% do exceptions.raise_compiler_error(missing_partition_msg) %}
         {% endif %}
-        
+
         {% set build_sql = bq_insert_overwrite(
             tmp_relation,
             target_relation,

plugins/snowflake/dbt/include/snowflake/macros/materializations/incremental.sql

@@ -27,7 +27,7 @@
 {% materialization incremental, adapter='snowflake' -%}
 
   {%- set unique_key = config.get('unique_key') -%}
-  {%- set full_refresh_mode = (flags.FULL_REFRESH == True) -%}
+  {%- set full_refresh_mode = (should_full_refresh()) -%}
 
   {% set target_relation = this %}
   {% set existing_relation = load_relation(this) %}

test/integration/005_simple_seed_test/models-downstream-seed/model.sql

@@ -0,0 +1 @@
+select * from {{ ref('seed_actual') }}

test/integration/005_simple_seed_test/test_simple_seed.py

@@ -7,7 +7,6 @@ class TestSimpleSeed(DBTIntegrationTest):
 
     def setUp(self):
         DBTIntegrationTest.setUp(self)
-
         self.run_sql_file("seed.sql")
 
     @property
@@ -16,7 +15,7 @@ def schema(self):
 
     @property
     def models(self):
-        return "models"
+        return "models-downstream-seed"
 
     @property
     def project_config(self):
@@ -28,27 +27,61 @@ def project_config(self):
             }
         }
 
+    def use_full_refresh_project(self, full_refresh: bool):
+        overrides = {
+            'seeds': {
+                'quote_columns': False,
+                'full_refresh': full_refresh,
+            }
+        }
+        self.use_default_project(overrides)
+
+    def _seed_and_run(self):
+        assert len(self.run_dbt(['seed'])) == 1
+        self.assertTablesEqual('seed_actual', 'seed_expected')
+
+        assert len(self.run_dbt(['run'])) == 1
+        self.assertTablesEqual('model', 'seed_expected')
+
+    def _after_seed_model_state(self, cmd, exists: bool):
+        assert len(self.run_dbt(cmd)) == 1
+        self.assertTablesEqual('seed_actual', 'seed_expected')
+        if exists:
+            self.assertTableDoesExist('model')
+        else:
+            self.assertTableDoesNotExist('model')
+
     @use_profile('postgres')
     def test_postgres_simple_seed(self):
-        results = self.run_dbt(["seed"])
-        self.assertEqual(len(results),  1)
-        self.assertTablesEqual("seed_actual","seed_expected")
+        self._seed_and_run()
 
         # this should truncate the seed_actual table, then re-insert.
-        results = self.run_dbt(["seed"])
-        self.assertEqual(len(results),  1)
-        self.assertTablesEqual("seed_actual","seed_expected")
+        self._after_seed_model_state(['seed'], exists=True)
 
     @use_profile('postgres')
-    def test_postgres_simple_seed_with_drop(self):
-        results = self.run_dbt(["seed"])
-        self.assertEqual(len(results),  1)
-        self.assertTablesEqual("seed_actual","seed_expected")
+    def test_postgres_simple_seed_full_refresh_flag(self):
+        self._seed_and_run()
 
-        # this should drop the seed table, then re-create
-        results = self.run_dbt(["seed", "--full-refresh"])
-        self.assertEqual(len(results),  1)
-        self.assertTablesEqual("seed_actual","seed_expected")
+        # this should drop the seed_actual table, then re-create it, so the
+        # model won't exist.
+        self._after_seed_model_state(['seed', '--full-refresh'], exists=False)
+
+    @use_profile('postgres')
+    def test_postgres_simple_seed_full_refresh_config(self):
+        self._seed_and_run()
+
+        # set the full_refresh config to False
+        self.use_full_refresh_project(False)
+
+        self._after_seed_model_state(['seed'], exists=True)
+        # make sure we ignore the full-refresh flag (the config is higher
+        # priority than the flag)
+        self._after_seed_model_state(['seed', '--full-refresh'], exists=True)
+
+        # this should drop the seed_actual table, then re-create it, so the
+        # model won't exist.
+        self.use_full_refresh_project(True)
+        self._after_seed_model_state(['seed'], exists=False)
 
 
 class TestSimpleSeedCustomSchema(DBTIntegrationTest):
@@ -89,7 +122,6 @@ def test_postgres_simple_seed_with_schema(self):
         self.assertEqual(len(results),  1)
         self.assertTablesEqual("seed_actual","seed_expected", table_a_schema=schema_name)
 
-
     @use_profile('postgres')
     def test_postgres_simple_seed_with_drop_and_schema(self):
         schema_name = "{}_{}".format(self.unique_schema(), 'custom_schema')

test/integration/012_deprecation_tests/models/already_exists.sql

@@ -1,5 +1,5 @@
 select 1 as id
 
-{% if adapter.already_exists(this.schema, this.identifier) and not flags.FULL_REFRESH %}
+{% if adapter.already_exists(this.schema, this.identifier) and not should_full_refresh() %}
 	where id > (select max(id) from {{this}})
 {% endif %}

test/integration/017_runtime_materialization_tests/test_runtime_materialization.py

@@ -25,10 +25,13 @@ def schema(self):
     def models(self):
         return "models"
 
+    def run_dbt_full_refresh(self):
+        return self.run_dbt(['run', '--full-refresh'])
+
     @use_profile('postgres')
     def test_postgres_full_refresh(self):
         # initial full-refresh should have no effect
-        results = self.run_dbt(['run', '--full-refresh'])
+        results = self.run_dbt_full_refresh()
         self.assertEqual(len(results), 3)
 
         self.assertTablesEqual("seed", "view")
@@ -37,13 +40,13 @@ def test_postgres_full_refresh(self):
 
         # adds one record to the incremental model. full-refresh should truncate then re-run
         self.run_sql_file("invalidate_incremental.sql")
-        results = self.run_dbt(['run', '--full-refresh'])
+        results = self.run_dbt_full_refresh()
         self.assertEqual(len(results), 3)
         self.assertTablesEqual("seed", "incremental")
 
         self.run_sql_file("update.sql")
 
-        results = self.run_dbt(['run', '--full-refresh'])
+        results = self.run_dbt_full_refresh()
         self.assertEqual(len(results), 3)
 
         self.assertTablesEqual("seed", "view")
@@ -59,3 +62,14 @@ def test_postgres_delete__dbt_tmp_relation(self):
 
         self.assertTableDoesNotExist('view__dbt_tmp')
         self.assertTablesEqual("seed", "view")
+
+
+class TestRuntimeMaterializationWithConfig(TestRuntimeMaterialization):
+    @property
+    def project_config(self):
+        result = super().project_config
+        result.update({'models': {'full_refresh': True}})
+        return result
+
+    def run_dbt_full_refresh(self):
+        return self.run_dbt(['run'])