def test_summarize_removes_1_grouping(backend):
data = data_frame(a=1, b=2, c=3)
df = backend.load_df(data)
q1 = df >> group_by(_.a, _.b) >> summarize(n=n(_))
assert q1.group_by == ("a")
q2 = q1 >> summarize(n=n(_))
assert not len(q2.group_by)
max_size=10))
OMNIBUS_VECTOR_FUNCS = [
#cumall, cumany, cummean,
#desc,
v.dense_rank(_.x, na_option="keep"),
#v.percent_rank(_.x),
v.min_rank(_.x, na_option="keep"),
v.cume_dist(_.x, na_option="keep"),
v.row_number(_.x),
#ntile,
v.between(_.x, 2, 5, default=False),
v.coalesce(_.x, 2),
v.lead(_.x),
v.lag(_.x),
v.n(_.x),
v.na_if(_.x, 2),
#near,
v.nth(_.x, 2),
v.first(_.x),
v.last(_.x,
order_by=_.x), # TODO: in SQL getting FROM LAST requires order by
]
VECTOR_AGG_FUNCS = [
v.n(_.x),
v.n(_),
]
VECTOR_FILTER_FUNCS = [
v.dense_rank(_.x, na_option="keep") < 2,
def test_summarize_unnamed_args(df):
assert_equal_query(df, summarize(n(_)), pd.DataFrame({'n(_)': 4}))
def test_summarize_removes_order_vars(backend, df):
lazy_tbl = df >> summarize(n=n(_))
assert not len(lazy_tbl.order_by)
def test_summarize_keeps_group_vars(backend, gdf):
q = gdf >> summarize(n=n(_))
assert list(q.last_op.c.keys()) == ["g", "n"]
def df(backend):
return backend.load_df(DATA)
@pytest.fixture(scope="module")
def df_float(backend):
return backend.load_df(DATA.assign(x=lambda d: d.x.astype(float)))
@pytest.fixture(scope="module")
def gdf(df):
return df >> group_by(_.g)
@pytest.mark.parametrize("query, output", [
(summarize(y=n(_)), data_frame(y=4)),
(summarize(y=_.x.min()), data_frame(y=1)),
])
def test_summarize_ungrouped(df, query, output):
assert_equal_query(df, query, output)
@pytest.mark.skip("TODO: should return 1 row (#63)")
def test_ungrouped_summarize_literal(df, query, output):
assert_equal_query(df, summarize(y=1), data_frame(y=1))
@backend_notimpl("sqlite")
def test_summarize_after_mutate_cuml_win(backend, df_float):
assert_equal_query(df_float,
mutate(y=_.x.cumsum()) >> summarize(z=_.y.max()),
def test_group_by_performs_mutate(df):
assert_equal_query(df,
group_by(z=_.x + _.y) >> summarize(n=n(_)),
data_frame(z=10, n=3))
def test_vector_n(x):
assert v.n(x) == 3
