| 631 | ]) |
| 632 | @pytest.mark.parametrize("from_pandas", [True, False]) |
| 633 | def test_sequence_numpy_double(seq, np_str_scalar, pa_type, from_pandas): |
| 634 | np_scalar = getattr(np, np_str_scalar) |
| 635 | data = [np_scalar(1.5), np_scalar(1), None, np_scalar(2.5), None, np.nan] |
| 636 | arr = pa.array(seq(data), from_pandas=from_pandas) |
| 637 | assert len(arr) == 6 |
| 638 | if from_pandas: |
| 639 | assert arr.null_count == 3 |
| 640 | else: |
| 641 | assert arr.null_count == 2 |
| 642 | if from_pandas: |
| 643 | # The NaN is skipped in type inference, otherwise it forces a |
| 644 | # float64 promotion |
| 645 | assert arr.type == pa_type |
| 646 | else: |
| 647 | assert arr.type == pa.float64() |
| 648 | |
| 649 | assert arr.to_pylist()[:4] == data[:4] |
| 650 | if from_pandas: |
| 651 | assert arr.to_pylist()[5] is None |
| 652 | else: |
| 653 | assert np.isnan(arr.to_pylist()[5]) |
| 654 | |
| 655 | |
| 656 | @pytest.mark.numpy |