AVRO-3827: [Rust] Disallow duplicate field names (#2433)

* AVRO-3827: [Rust] Disallow duplicate field names * Reuse BTreeMap. * Use the returned Option.
apache · Aug 10, 2023 · 3deb73c · 3deb73c
1 parent 7b50d06
commit 3deb73c
Show file tree

Hide file tree

Showing 2 changed files with 73 additions and 1 deletion.
diff --git a/lang/rust/avro/src/error.rs b/lang/rust/avro/src/error.rs
@@ -310,6 +310,9 @@ pub enum Error {
     #[error("Invalid field name {0}")]
     FieldName(String),
 
+    #[error("Duplicate field name {0}")]
+    FieldNameDuplicate(String),
+
     #[error("Invalid schema name {0}. It must match the regex '{1}'")]
     InvalidSchemaName(String, &'static str),
 

diff --git a/lang/rust/avro/src/schema.rs b/lang/rust/avro/src/schema.rs
@@ -1441,7 +1441,9 @@ impl Parser {
             })?;
 
         for field in &fields {
-            lookup.insert(field.name.clone(), field.position);
+            if let Some(_old) = lookup.insert(field.name.clone(), field.position) {
+                return Err(Error::FieldNameDuplicate(field.name.clone()));
+            }
 
             if let Some(ref field_aliases) = field.aliases {
                 for alias in field_aliases {
@@ -5064,4 +5066,71 @@ mod tests {
             other => Err(format!("Expected Error::FieldName, got {other:?}").into()),
         }
     }
+
+    #[test]
+    fn test_avro_3827_disallow_duplicate_field_names() -> TestResult {
+        let schema_str = r#"
+        {
+          "name": "my_schema",
+          "type": "record",
+          "fields": [
+            {
+              "name": "f1",
+              "type": {
+                "name": "a",
+                "type": "record",
+                "fields": []
+              }
+            },  {
+              "name": "f1",
+              "type": {
+                "name": "b",
+                "type": "record",
+                "fields": []
+              }
+            }
+          ]
+        }
+        "#;
+
+        match Schema::parse_str(schema_str) {
+            Err(Error::FieldNameDuplicate(_)) => (),
+            other => {
+                return Err(format!("Expected Error::FieldNameDuplicate, got {other:?}").into())
+            }
+        };
+
+        let schema_str = r#"
+        {
+          "name": "my_schema",
+          "type": "record",
+          "fields": [
+            {
+              "name": "f1",
+              "type": {
+                "name": "a",
+                "type": "record",
+                "fields": [
+                  {
+                    "name": "f1",
+                    "type": {
+                      "name": "b",
+                      "type": "record",
+                      "fields": []
+                    }
+                  }
+                ]
+              }
+            }
+          ]
+        }
+        "#;
+
+        let expected = r#"{"name":"my_schema","type":"record","fields":[{"name":"f1","type":{"name":"a","type":"record","fields":[{"name":"f1","type":{"name":"b","type":"record","fields":[]}}]}}]}"#;
+        let schema = Schema::parse_str(schema_str)?;
+        let canonical_form = schema.canonical_form();
+        assert_eq!(canonical_form, expected);
+
+        Ok(())
+    }
 }