Use safe_merge to handle multiple definitions of a single field

CHANGELOG.next.md

@@ -58,6 +58,7 @@ Thanks, you're awesome :-) -->
 * Add support for reusing offical fieldsets in custom schemas. #751
 * Add full path names to reused fieldsets in `nestings` array in `ecs_nested.yml`. #803
 * Allow shorthand notation for including all subfields in subsets. #805
+* Prevent custom schema from defining a type more than once. #821
 
 #### Deprecated
 

scripts/requirements.txt

@@ -2,3 +2,4 @@ pip
 PyYAML==5.3b1
 autopep8==1.4.4
 yamllint==1.19.0
+mock

scripts/schema_reader.py

@@ -28,7 +28,7 @@ def load_schema_files(files):
     fields_nested = {}
     for f in files:
         new_fields = read_schema_file(f)
-        fields_nested.update(new_fields)
+        fields_nested = ecs_helpers.safe_merge_dicts(fields_nested, new_fields)
     return fields_nested
 
 

scripts/tests/test_schema_reader.py

@@ -1,6 +1,7 @@
 import os
 import sys
 import unittest
+import mock
 
 sys.path.append(os.path.join(os.path.dirname(__file__), '..'))
 
@@ -71,6 +72,47 @@ def test_load_schemas_with_empty_list_loads_nothing(self):
         result = schema_reader.load_schemas([])
         self.assertEqual(result, ({}))
 
+    @mock.patch('scripts.schema_reader.read_schema_file')
+    def test_load_schemas_fail_on_redefinition(self, mock_read_schema):
+        mock_read_schema.side_effect = [
+            {
+                'file': {
+                    'name': 'file',
+                    'type': 'keyword'
+                }
+            },
+            {
+                'file': {
+                    'name': 'file',
+                    'type': 'text'
+                }
+            }
+        ]
+        with self.assertRaises(ValueError):
+            schema_reader.load_schema_files(['a', 'b'])
+
+    @mock.patch('scripts.schema_reader.read_schema_file')
+    def test_load_schemas_allows_unique_fields(self, mock_read_schema):
+        file_map = {
+            'file': {
+                'name': 'file',
+                'type': 'keyword'
+            }
+        }
+        host_map = {
+            'host': {
+                'name': 'host',
+                'type': 'text'
+            }
+        }
+        mock_read_schema.side_effect = [file_map, host_map]
+        exp = {
+            'file': file_map['file'],
+            'host': host_map['host']
+        }
+        res = schema_reader.load_schema_files(['a', 'b'])
+        self.assertEqual(res, exp)
+
     def test_flatten_fields(self):
         fields = {
             'top_level': {