def test_udfs(self, source_tbl, udf_name, udf, is_python_udf, query,
expected_result):
if is_python_udf:
register_python(udf_name, udf)
else:
register_numpy(udf_name, udf)
actual_tbl = source_tbl.sql(query)
_assert_tables_equal(actual_tbl, expected_result)
def test_update_udf(self, source_tbl, is_python_udf):
def f_square(x):
return x**2
def f_cube(x):
return x**3
if is_python_udf:
register_python('udf_upd', f_square)
register_python('udf_upd', f_cube)
else:
register_python('udf_upd', f_square)
register_python('udf_upd', f_cube)
actual_tbl = source_tbl.sql(
"select udf_upd(id) as pow from t order by pow")
expected = {
'pow': (1, 8, 27, 64, 125, 216, 343, 512),
}
_assert_tables_equal(actual_tbl, expected)
def test_from_pandas(self, input, query, expected_result):
pdf = pd.DataFrame(data=input)
tbl = Table.from_pandas(pdf)
actual_tbl = tbl.sql(query)
_assert_tables_equal(actual_tbl, expected_result)
def test_from_pyarrow(self, input, query, expected_result):
arrow_table = pa.Table.from_pydict(input)
tbl = Table.from_arrow(arrow_table)
actual_tbl = tbl.sql(query)
_assert_tables_equal(actual_tbl, expected_result)
def test_from_pydict(self, input, query, expected_result):
actual_tbl = Table.from_pydict(input).sql(query)
_assert_tables_equal(actual_tbl, expected_result)
def test_read_parquet(self, query, expected_result, parquet_datafile):
actual_tbl = read_parquet(parquet_datafile).sql(query)
_assert_tables_equal(actual_tbl, expected_result)
def test_open_csv(self, query, expected_result, csv_datafile):
actual_tbl = stream_csv(csv_datafile).sql(query)
_assert_tables_equal(actual_tbl, expected_result)
def test_queries(self, source_tbl, query, expected_result):
actual_tbl = source_tbl.sql(query)
_assert_tables_equal(actual_tbl, expected_result)