override seed types

dbt/adapters/bigquery.py

@@ -484,7 +484,8 @@ def convert_datetime_type(cls, agate_table, col_idx):
         return "datetime"
 
     @classmethod
-    def create_csv_table(cls, profile, schema, table_name, agate_table):
+    def create_csv_table(cls, profile, schema, table_name, agate_table,
+                         column_override):
         pass
 
     @classmethod
@@ -493,17 +494,19 @@ def reset_csv_table(cls, profile, schema, table_name, agate_table,
         cls.drop(profile, schema, table_name, "table")
 
     @classmethod
-    def _agate_to_schema(cls, agate_table):
+    def _agate_to_schema(cls, agate_table, column_override):
         bq_schema = []
         for idx, col_name in enumerate(agate_table.column_names):
-            type_ = cls.convert_agate_type(agate_table, idx)
+            inferred_type = cls.convert_agate_type(agate_table, idx)
+            type_ = column_override.get(col_name, inferred_type)
             bq_schema.append(
                 google.cloud.bigquery.SchemaField(col_name, type_))
         return bq_schema
 
     @classmethod
-    def load_csv_rows(cls, profile, schema, table_name, agate_table):
-        bq_schema = cls._agate_to_schema(agate_table)
+    def load_csv_rows(cls, profile, schema, table_name, agate_table,
+                      column_override):
+        bq_schema = cls._agate_to_schema(agate_table, column_override)
         dataset = cls.get_dataset(profile, schema, None)
         table = dataset.table(table_name, schema=bq_schema)
         conn = cls.get_connection(profile, None)

dbt/adapters/default.py

@@ -99,7 +99,8 @@ def cancel_connection(cls, project, connection):
             '`cancel_connection` is not implemented for this adapter!')
 
     @classmethod
-    def create_csv_table(cls, profile, schema, table_name, agate_table):
+    def create_csv_table(cls, profile, schema, table_name, agate_table,
+                         column_override):
         raise dbt.exceptions.NotImplementedException(
             '`create_csv_table` is not implemented for this adapter!')
 
@@ -110,7 +111,8 @@ def reset_csv_table(cls, profile, schema, table_name, agate_table,
             '`reset_csv_table` is not implemented for this adapter!')
 
     @classmethod
-    def load_csv_rows(cls, profile, schema, table_name, agate_table):
+    def load_csv_rows(cls, profile, schema, table_name, agate_table,
+                      column_override):
         raise dbt.exceptions.NotImplementedException(
             '`load_csv_rows` is not implemented for this adapter!')
 
@@ -651,18 +653,20 @@ def quote_schema_and_table(cls, profile, schema, table, model_name=None):
 
     @classmethod
     def handle_csv_table(cls, profile, schema, table_name, agate_table,
-                         full_refresh=False):
+                         column_override, full_refresh=False):
         existing = cls.query_for_existing(profile, schema)
         existing_type = existing.get(table_name)
         if existing_type and existing_type != "table":
             raise dbt.exceptions.RuntimeException(
                 "Cannot seed to '{}', it is a view".format(table_name))
         if existing_type:
             cls.reset_csv_table(profile, schema, table_name, agate_table,
-                                full_refresh=full_refresh)
+                                column_override, full_refresh=full_refresh)
         else:
-            cls.create_csv_table(profile, schema, table_name, agate_table)
-        cls.load_csv_rows(profile, schema, table_name, agate_table)
+            cls.create_csv_table(profile, schema, table_name, agate_table,
+                                 column_override)
+        cls.load_csv_rows(profile, schema, table_name, agate_table,
+                          column_override)
         cls.commit_if_has_connection(profile, None)
 
     @classmethod

dbt/adapters/postgres.py

@@ -194,26 +194,32 @@ def convert_time_type(cls, agate_table, col_idx):
         return "time"
 
     @classmethod
-    def create_csv_table(cls, profile, schema, table_name, agate_table):
+    def create_csv_table(cls, profile, schema, table_name, agate_table,
+                         column_override):
         col_sqls = []
         for idx, col_name in enumerate(agate_table.column_names):
-            type_ = cls.convert_agate_type(agate_table, idx)
+            inferred_type = cls.convert_agate_type(agate_table, idx)
+            type_ = column_override.get(col_name, inferred_type)
             col_sqls.append('{} {}'.format(col_name, type_))
         sql = 'create table "{}"."{}" ({})'.format(schema, table_name,
                                                    ", ".join(col_sqls))
         return cls.add_query(profile, sql)
 
     @classmethod
     def reset_csv_table(cls, profile, schema, table_name, agate_table,
-                        full_refresh=False):
+                        column_override, full_refresh=False):
         if full_refresh:
             cls.drop_table(profile, schema, table_name, None)
-            cls.create_csv_table(profile, schema, table_name, agate_table)
+            cls.create_csv_table(profile, schema, table_name, agate_table,
+                                 column_override)
         else:
             cls.truncate(profile, schema, table_name)
 
     @classmethod
-    def load_csv_rows(cls, profile, schema, table_name, agate_table):
+    def load_csv_rows(cls, profile, schema, table_name, agate_table,
+                      column_override):
+        bindings = []
+        placeholders = []
         cols_sql = ", ".join(c for c in agate_table.column_names)
 
         for chunk in chunks(agate_table.rows, 10000):

dbt/contracts/graph/parsed.py

@@ -25,6 +25,7 @@
     Required('post-hook'): [hook_contract],
     Required('pre-hook'): [hook_contract],
     Required('vars'): dict,
+    Required('column_types'): dict,
 }, extra=ALLOW_EXTRA)
 
 parsed_node_contract = unparsed_node_contract.extend({

dbt/model.py

@@ -12,7 +12,7 @@ class SourceConfig(object):
     ConfigKeys = DBTConfigKeys
 
     AppendListFields = ['pre-hook', 'post-hook']
-    ExtendDictFields = ['vars']
+    ExtendDictFields = ['vars', 'column_types']
     ClobberFields = [
         'schema',
         'enabled',

dbt/node_runners.py

@@ -509,7 +509,10 @@ def execute(self, compiled_node, existing_, flat_graph):
         schema = compiled_node["schema"]
         table_name = compiled_node["name"]
         table = compiled_node["agate_table"]
+
+        column_override = compiled_node['config'].get('column_types', {})
         self.adapter.handle_csv_table(self.profile, schema, table_name, table,
+                                      column_override,
                                       full_refresh=dbt.flags.FULL_REFRESH)
 
         if dbt.flags.FULL_REFRESH:

dbt/utils.py

@@ -26,6 +26,7 @@
     'pre-hook',
     'post-hook',
     'vars',
+    'column_types',
     'bind',
 ]
 

test/integration/005_simple_seed_test/macros/schema_test.sql

@@ -0,0 +1,15 @@
+
+{% macro test_column_type(model, field, type) %}
+
+    {% set cols = adapter.get_columns_in_table(model.schema, model.name) %}
+
+    {% set col_types = {} %}
+    {% for col in cols %}
+        {% set _ = col_types.update({col.name: col.data_type}) %}
+    {% endfor %}
+
+    {% set val = 0 if col_types[field] == type else 1 %}
+
+    select {{ val }} as pass_fail
+
+{% endmacro %}

test/integration/005_simple_seed_test/models/schema.yml

@@ -0,0 +1,7 @@
+
+
+seed_enabled:
+    constraints:
+        column_type:
+            - {field: id, type: 'character varying(255)' }
+            - {field: birthday, type: 'date' }

test/integration/005_simple_seed_test/test_simple_seed.py

@@ -121,3 +121,37 @@ def test_simple_seed_with_disabled(self):
         self.assertTableDoesExist('seed_enabled')
         self.assertTableDoesNotExist('seed_disabled')
 
+
+class TestSimpleSeedColumnOverride(DBTIntegrationTest):
+
+    @property
+    def schema(self):
+        return "simple_seed_005"
+
+    @property
+    def models(self):
+        return "test/integration/005_simple_seed_test/models"
+
+    @property
+    def project_config(self):
+        return {
+            "data-paths": ['test/integration/005_simple_seed_test/data-config'],
+            "macro-paths": ['test/integration/005_simple_seed_test/macros'],
+            "seeds": {
+                "test": {
+                    "enabled": False,
+                    "seed_enabled": {
+                        "enabled": True,
+                        "column_types": {
+                            "id": "text",
+                            "birthday": "date",
+                        }
+                    },
+                }
+            }
+        }
+
+    @attr(type='postgres')
+    def test_simple_seed_with_column_override(self):
+        self.run_dbt(["seed"])
+        self.run_dbt(["test"])

test/unit/test_compiler.py

@@ -38,6 +38,7 @@ def setUp(self):
             'post-hook': [],
             'pre-hook': [],
             'vars': {},
+            'column_types': {},
         }
 
     def test__prepend_ctes__already_has_cte(self):

test/unit/test_parser.py

@@ -57,6 +57,7 @@ def setUp(self):
             'post-hook': [],
             'pre-hook': [],
             'vars': {},
+            'column_types': {},
         }
 
         self.disabled_config = {
@@ -65,6 +66,7 @@ def setUp(self):
             'post-hook': [],
             'pre-hook': [],
             'vars': {},
+            'column_types': {},
         }
 
     def test__single_model(self):