Set default to error instead of drop (#259)

when encountering bad input data, we can default to principle of least
surprise and raise an exception.

Co-authored-by: Chang She <chang@lancedb.com>

python/tests/test_table.py

@@ -181,7 +181,21 @@ def test_create_index_method():
 
 
 def test_add_with_nans(db):
-    # By default we drop bad input vectors
+    # by default we raise an error on bad input vectors
+    bad_data = [
+        {"vector": [np.nan], "item": "bar", "price": 20.0},
+        {"vector": [5], "item": "bar", "price": 20.0},
+        {"vector": [np.nan, np.nan], "item": "bar", "price": 20.0},
+        {"vector": [np.nan, 5.0], "item": "bar", "price": 20.0},
+    ]
+    for row in bad_data:
+        with pytest.raises(ValueError):
+            LanceTable.create(
+                db,
+                "error_test",
+                data=[{"vector": [3.1, 4.1], "item": "foo", "price": 10.0}, row],
+            )
+
     table = LanceTable.create(
         db,
         "drop_test",
@@ -191,6 +205,7 @@ def test_add_with_nans(db):
             {"vector": [5], "item": "bar", "price": 20.0},
             {"vector": [np.nan, np.nan], "item": "bar", "price": 20.0},
         ],
+        on_bad_vectors="drop",
     )
     assert len(table) == 1
 
@@ -210,18 +225,3 @@ def test_add_with_nans(db):
     arrow_tbl = table.to_lance().to_table(filter="item == 'bar'")
     v = arrow_tbl["vector"].to_pylist()[0]
     assert np.allclose(v, np.array([0.0, 0.0]))
-
-    bad_data = [
-        {"vector": [np.nan], "item": "bar", "price": 20.0},
-        {"vector": [5], "item": "bar", "price": 20.0},
-        {"vector": [np.nan, np.nan], "item": "bar", "price": 20.0},
-        {"vector": [np.nan, 5.0], "item": "bar", "price": 20.0},
-    ]
-    for row in bad_data:
-        with pytest.raises(ValueError):
-            LanceTable.create(
-                db,
-                "raise_test",
-                data=[{"vector": [3.1, 4.1], "item": "foo", "price": 10.0}, row],
-                on_bad_vectors="raise",
-            )