def test_auto_blocksize_csv(monkeypatch): psutil = pytest.importorskip("psutil") total_memory = psutil.virtual_memory().total cpu_count = psutil.cpu_count() mock_read_bytes = mock.Mock(wraps=read_bytes) monkeypatch.setattr(dask.dataframe.io.csv, "read_bytes", mock_read_bytes) expected_block_size = auto_blocksize(total_memory, cpu_count) with filetexts(csv_files, mode="b"): dd.read_csv("2014-01-01.csv") assert mock_read_bytes.called assert mock_read_bytes.call_args[1]["blocksize"] == expected_block_size
def test_auto_blocksize_csv(monkeypatch): psutil = pytest.importorskip('psutil') try: from unittest import mock except ImportError: mock = pytest.importorskip('mock') total_memory = psutil.virtual_memory().total cpu_count = psutil.cpu_count() mock_read_bytes = mock.Mock(wraps=read_bytes) monkeypatch.setattr(dask.dataframe.io.csv, 'read_bytes', mock_read_bytes) expected_block_size = auto_blocksize(total_memory, cpu_count) with filetexts(csv_files, mode='b'): dd.read_csv('2014-01-01.csv') assert mock_read_bytes.called assert mock_read_bytes.call_args[1]['blocksize'] == expected_block_size
def test_auto_blocksize_max64mb(): blocksize = auto_blocksize(1000000000000, 3) assert blocksize == int(64e6) assert isinstance(blocksize, int)
def test_auto_blocksize(): assert isinstance(auto_blocksize(3000, 15), int) assert auto_blocksize(3000, 3) == 100 assert auto_blocksize(5000, 2) == 250