fix(rust, python): fix ndjson empty array parsing (#6785)

pola-rs · Feb 10, 2023 · 729dece · 729dece
1 parent 456d704
commit 729dece
Show file tree

Hide file tree

Showing 2 changed files with 12 additions and 2 deletions.
diff --git a/polars/polars-io/src/ndjson_core/buffer.rs b/polars/polars-io/src/ndjson_core/buffer.rs
@@ -168,9 +168,13 @@ fn value_to_dtype(val: &Value) -> DataType {
         Value::Static(StaticNode::F64(_)) => DataType::Float64,
         Value::Static(StaticNode::Null) => DataType::Null,
         Value::Array(arr) => {
-            let dtype = value_to_dtype(&arr[0]);
+            let inner_type = if let Some(val) = arr.first() {
+                value_to_dtype(val)
+            } else {
+                DataType::Unknown
+            };
 
-            DataType::List(Box::new(dtype))
+            DataType::List(Box::new(inner_type))
         }
         #[cfg(feature = "dtype-struct")]
         Value::Object(doc) => {

diff --git a/py-polars/tests/unit/io/test_json.py b/py-polars/tests/unit/io/test_json.py
@@ -83,3 +83,9 @@ def test_write_ndjson_with_trailing_newline() -> None:
 
     expected = pl.DataFrame({"Column1": ["Value1"]})
     assert_frame_equal(df, expected)
+
+
+def test_read_ndjson_empty_array() -> None:
+    assert pl.read_ndjson(io.StringIO("""{"foo": {"bar": []}}""")).to_dict(False) == {
+        "foo": [{"": None}]
+    }