Fix: Support nullable union types in Avro array items

CHANGELOG.md

@@ -39,6 +39,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 - Improved regex for extracting Azure storage account names from URLs with containerName@storageAccountName format (#848)
 - JSON Schema Check: Add globbing support for local JSON files
+- Avro Importer: Optional and required enum types now work for arrays
 - Fixed server section rendering for markdown exporter
 
 ## [0.10.34] - 2025-08-06

datacontract/imports/avro_importer.py

@@ -194,6 +194,8 @@ def import_avro_array_items(array_schema: avro.schema.ArraySchema) -> Field:
     elif array_schema.items.type == "array":
         items.type = "array"
         items.items = import_avro_array_items(array_schema.items)
+    elif array_schema.items.type == "union":
+        items.type = import_type_from_union_schema(array_schema.items)
     else:  # primitive type
         items.type = map_type_from_avro(array_schema.items.type)
 
@@ -227,25 +229,28 @@ def import_avro_map_values(map_schema: avro.schema.MapSchema) -> Field:
 
 
 def import_type_of_optional_field(field: avro.schema.Field) -> str:
+    return import_type_from_union_schema(field.type)
+
+def import_type_from_union_schema(union_schema: avro.schema.UnionSchema) -> str:
     """
-    Determine the type of optional field in an Avro union.
+    Extract the first non-null type from a union schema and map it to data contract type.
 
     Args:
-        field: The Avro field with a union type.
+        union_schema (avro.schema.UnionSchema): The Avro union schema.
 
     Returns:
-        str: The mapped type of the non-null field in the union.
+        str: The mapped type of the first non-null field in the union.
 
     Raises:
         DataContractException: If no non-null type is found in the union.
     """
-    for field_type in field.type.schemas:
-        if field_type.type != "null":
-            logical_type = field_type.get_prop("logicalType")
+    for schema_type in union_schema.schemas:
+        if schema_type.type != "null":
+            logical_type = schema_type.get_prop("logicalType")
             if logical_type and logical_type in LOGICAL_TYPE_MAPPING:
                 return LOGICAL_TYPE_MAPPING[logical_type]
             else:
-                return map_type_from_avro(field_type.type)
+                return map_type_from_avro(schema_type.type)
     raise DataContractException(
         type="schema",
         result="failed",

tests/fixtures/avro/data/non_nullable_union.avsc

@@ -0,0 +1,17 @@
+{
+  "type": "record",
+  "name": "TestNonNullableUnion",
+  "doc": "Test record with non-nullable union types",
+  "fields": [
+    {
+      "name": "id",
+      "type": "string",
+      "doc": "Required field"
+    },
+    {
+      "name": "multi_type_field",
+      "type": ["string", "int", "boolean"],
+      "doc": "Field with multiple types takes first non-null type"
+    }
+  ]
+}

tests/fixtures/avro/data/nullable_array_items.avsc

@@ -0,0 +1,33 @@
+{
+  "type": "record",
+  "name": "TestNullableArrayItems",
+  "doc": "Test record with nullable unions in array items",
+  "fields": [
+    {
+      "name": "id",
+      "type": "string",
+      "doc": "Required field - works fine"
+    },
+    {
+      "name": "nullable_string", 
+      "type": ["null", "string"],
+      "doc": "Simple nullable field",
+      "default": null
+    },
+    {
+      "name": "media_urls",
+      "type": [
+        "null",
+        {
+          "type": "array",
+          "items": [
+            "null",
+            "string"
+          ]
+        }
+      ],
+      "default": null,
+      "doc": "Nullable array with nullable string items "
+    }
+  ]
+}

tests/fixtures/avro/export/datacontract_non_nullable_union.avsc

@@ -0,0 +1,20 @@
+{
+  "type": "record",
+  "name": "TestNonNullableUnion",
+  "doc": "Test record with non-nullable union types",
+  "fields": [
+    {
+      "name": "id",
+      "doc": "Required field",
+      "type": "string"
+    },
+    {
+      "name": "multi_type_field",
+      "doc": "Field with multiple types takes first non-null type",
+      "type": [
+        "null",
+        "string"
+      ]
+    }
+  ]
+}

tests/fixtures/avro/export/datacontract_non_nullable_union.yaml

@@ -0,0 +1,17 @@
+dataContractSpecification: 1.2.0
+id: non-nullable-union
+info:
+  title: Test Non-Nullable Union
+  version: 0.0.1
+models:
+  TestNonNullableUnion:
+    description: Test record with non-nullable union types
+    fields:
+      id:
+        type: string
+        required: true
+        description: Required field
+      multi_type_field:
+        type: string
+        required: false
+        description: Field with multiple types takes first non-null type

tests/fixtures/avro/export/datacontract_nullable_array_items.avsc

@@ -0,0 +1,31 @@
+{
+  "type": "record",
+  "name": "TestNullableArrayItems",
+  "doc": "Test record with nullable unions in array items",
+  "fields": [
+    {
+      "name": "id",
+      "doc": "Required field",
+      "type": "string"
+    },
+    {
+      "name": "nullable_string",
+      "doc": "Simple nullable fields",
+      "type": [
+        "null",
+        "string"
+      ]
+    },
+    {
+      "name": "media_urls",
+      "doc": "Nullable array with nullable string items",
+      "type": [
+        "null",
+        {
+          "type": "array",
+          "items": "string"
+        }
+      ]
+    }
+  ]
+}

tests/fixtures/avro/export/datacontract_nullable_array_items.yaml

@@ -0,0 +1,23 @@
+dataContractSpecification: 1.2.0
+id: nullable-array-items
+info:
+  title: Test Nullable Array Items
+  version: 0.0.1
+models:
+  TestNullableArrayItems:
+    description: Test record with nullable unions in array items
+    fields:
+      id:
+        type: string
+        required: true
+        description: Required field
+      nullable_string:
+        type: string
+        required: false
+        description: Simple nullable fields
+      media_urls:
+        type: array
+        required: false
+        description: Nullable array with nullable string items
+        items:
+          type: string

tests/test_export_avro.py

@@ -101,3 +101,25 @@ def test_to_field_float():
     result = to_avro_schema_json(model_name, model)
 
     assert json.loads(result) == json.loads(expected_avro_schema)
+
+
+def test_export_avro_nullable_array_items():
+    data_contract = DataContractSpecification.from_file("fixtures/avro/export/datacontract_nullable_array_items.yaml")
+    with open("fixtures/avro/export/datacontract_nullable_array_items.avsc") as file:
+        expected_avro_schema = file.read()
+
+    model_name, model = next(iter(data_contract.models.items()))
+    result = to_avro_schema_json(model_name, model)
+
+    assert json.loads(result) == json.loads(expected_avro_schema)
+
+
+def test_export_avro_non_nullable_union():
+    data_contract = DataContractSpecification.from_file("fixtures/avro/export/datacontract_non_nullable_union.yaml")
+    with open("fixtures/avro/export/datacontract_non_nullable_union.avsc") as file:
+        expected_avro_schema = file.read()
+
+    model_name, model = next(iter(data_contract.models.items()))
+    result = to_avro_schema_json(model_name, model)
+
+    assert json.loads(result) == json.loads(expected_avro_schema)

tests/test_import_avro.py

@@ -352,3 +352,63 @@ def test_import_avro_optional_enum():
     print("Result:\n", result.to_yaml())
     assert yaml.safe_load(result.to_yaml()) == yaml.safe_load(expected)
     assert DataContract(data_contract_str=expected).lint().has_passed()
+
+
+def test_import_avro_nullable_array_items():
+    result = DataContract().import_from_source("avro", "fixtures/avro/data/nullable_array_items.avsc")
+
+    expected = """
+dataContractSpecification: 1.2.0
+id: my-data-contract-id
+info:
+  title: My Data Contract
+  version: 0.0.1
+models:
+  TestNullableArrayItems:
+    description: Test record with nullable unions in array items
+    fields:
+      id:
+        type: string
+        required: true
+        description: Required field - works fine
+      nullable_string:
+        type: string
+        required: false
+        description: Simple nullable field
+      media_urls:
+        type: array
+        required: false
+        description: 'Nullable array with nullable string items '
+        items:
+          type: string
+"""
+    print("Result:\n", result.to_yaml())
+    assert yaml.safe_load(result.to_yaml()) == yaml.safe_load(expected)
+    assert DataContract(data_contract_str=expected).lint(enabled_linters="none").has_passed()
+
+
+def test_import_avro_non_nullable_union():
+    result = DataContract().import_from_source("avro", "fixtures/avro/data/non_nullable_union.avsc")
+
+    expected = """
+dataContractSpecification: 1.2.0
+id: my-data-contract-id
+info:
+  title: My Data Contract
+  version: 0.0.1
+models:
+  TestNonNullableUnion:
+    description: Test record with non-nullable union types
+    fields:
+      id:
+        type: string
+        required: true
+        description: Required field
+      multi_type_field:
+        type: string
+        required: false
+        description: Field with multiple types takes first non-null type
+"""
+    print("Result:\n", result.to_yaml())
+    assert yaml.safe_load(result.to_yaml()) == yaml.safe_load(expected)
+    assert DataContract(data_contract_str=expected).lint().has_passed()