fix: set doc id first (#175)

numb3r3 · web-flow · commit 50722a82e85b · 2022-03-07T09:22:53.000+01:00
diff --git a/docarray/base.py b/docarray/base.py
@@ -46,6 +46,11 @@ def __init__(
                     kwargs.pop(k)
 
             self._data = self._data_class(self)
+
+            # ``id`` need to be set at the first place, since some other attributes would depends on it.
+            if 'id' in kwargs:
+                setattr(self._data, 'id', kwargs.pop('id'))
+
             for k, v in kwargs.items():
                 setattr(self._data, k, v)
 
diff --git a/tests/unit/document/test_docdata.py b/tests/unit/document/test_docdata.py
@@ -23,6 +23,14 @@ def test_doc_hash_identical():
     assert d1 == d2
 
 
+def test_doc_id_setter():
+    d = Document(
+        chunks=[Document(id='c0')],
+        id='d0',
+    )
+    assert d.id == d.chunks[0].parent_id == 'd0'
+
+
 def test_doc_hash_complicate_content():
     d1 = Document(text='hello', embedding=np.array([1, 2, 3]), id=1)
     d2 = Document(text='hello', embedding=np.array([1, 2, 3]), id=1)
