Override data type when copying variable (#392)

* Failing test

* Override the data type when copying a variable

* Fix mypy

Author: mmwinther

src/dapla_metadata/datasets/core.py

@@ -561,22 +561,21 @@ def copy_variable(
 
         source_short_name = source_short_name or target_short_name
 
-        metadata_document_variables = read_variables_from_metadata_document(
-            metadata_document_path
-        )
+        source_variables = read_variables_from_metadata_document(metadata_document_path)
 
-        variables_by_short_name = {
-            v.short_name: v
-            for v in metadata_document_variables
-            if v.short_name is not None
+        source_variables_lookup = {
+            v.short_name: v for v in source_variables if v.short_name is not None
         }
 
-        if source_short_name not in variables_by_short_name:
+        if source_short_name not in source_variables_lookup:
             msg = f"{source_short_name} does not exist!"
             raise ValueError(msg)
 
-        source_variable = variables_by_short_name[source_short_name]
-
+        source_variable = source_variables_lookup[source_short_name]
+        # Always override the data type to ensure it matches the physical dataset.
+        source_variable.data_type = self.variables_lookup[  # type: ignore[assignment]
+            target_short_name
+        ].data_type
         self.variables_lookup[target_short_name] = source_variable
 
         source_variable = all_optional_model.Variable.model_validate(source_variable)

tests/datasets/resources/metadata_documents/copy_variables.json

@@ -153,7 +153,7 @@
                         "languageText": "Ny persid"
                     }
                 ],
-                "data_type": "STRING",
+                "data_type": "ARRAY",
                 "variable_role": "IDENTIFIER",
                 "definition_uri": "https://www.ssb.no/a/metadata/conceptvariable/vardok/26/nb",
                 "is_personal_data": true,

tests/datasets/test_copy_variables.py

@@ -25,13 +25,15 @@ def test_copy_variable(
 
     with pathlib.Path.open(written_document) as f:
         written_metadata = json.loads(f.read())
-        datadoc_metadata = written_metadata["datadoc"]["variables"]
 
-        assert datadoc_metadata[0]["short_name"] == target_short_name
-        correct_variable = datadoc_metadata[0]
-        assert correct_variable["name"] is not None
-        assert correct_variable["name"][2]["languageCode"] == "nb"
-        assert correct_variable["name"][2]["languageText"] == "Ny persid"
+    variables = written_metadata["datadoc"]["variables"]
+    assert variables[0]["short_name"] == target_short_name
+    correct_variable = variables[0]
+    assert correct_variable["name"] is not None
+    assert correct_variable["name"][2]["languageCode"] == "nb"
+    assert correct_variable["name"][2]["languageText"] == "Ny persid"
+    # Data type must match that from the dataset.
+    assert correct_variable["data_type"] == "STRING"
 
 
 def test_variable_not_in_target_dataset(