def test_summary_on_ndarray():
assert compute(summary(total=a.sum(), min=a.min()), ax) == (ax.min(), ax.sum())
result = compute(summary(total=a.sum(), min=a.min(), keepdims=True), ax)
expected = np.array([(ax.min(), ax.sum())], dtype=[("min", "float32"), ("total", "float64")])
assert result.ndim == ax.ndim
assert eq(expected, result)
def test_summary_axis():
x = symbol('x', '5 * 3 * float32')
assert summary(a=x.min(), b=x.max(), axis=0).dshape == \
dshape('3 * {a: float32, b: float32}')
assert summary(a=x.min(), b=x.max(), axis=1).dshape == \
dshape('5 * {a: float32, b: float32}')
assert summary(a=x.min(), b=x.max(), axis=1, keepdims=True).dshape == \
dshape('5 * 1 * {a: float32, b: float32}')
def test_summary_on_series():
ser = Series([1, 2, 3])
s = symbol('s', '3 * int')
expr = summary(max=s.max(), min=s.min())
assert compute(expr, ser) == (3, 1)
expr = summary(max=s.max(), min=s.min(), keepdims=True)
assert compute(expr, ser) == [(3, 1)]
def test_summary_on_series():
ser = dd.from_pandas(pd.Series([1, 2, 3]), npartitions=2)
s = symbol('s', '3 * int')
expr = summary(max=s.max(), min=s.min())
assert compute(expr, ser) == (3, 1)
expr = summary(max=s.max(), min=s.min(), keepdims=True)
assert compute(expr, ser) == [(3, 1)]
def test_summary_on_series():
ser = dd.from_pandas(pd.Series([1, 2, 3]), npartitions=2)
s = symbol('s', '3 * int')
expr = summary(max=s.max(), min=s.min())
assert compute(expr, ser) == (3, 1)
expr = summary(max=s.max(), min=s.min(), keepdims=True)
assert compute(expr, ser) == [(3, 1)]
def test_summary_on_series():
ser = Series([1, 2, 3])
s = symbol('s', '3 * int')
expr = summary(max=s.max(), min=s.min())
assert compute(expr, ser) == (3, 1)
expr = summary(max=s.max(), min=s.min(), keepdims=True)
assert compute(expr, ser) == [(3, 1)]
def test_summary_axis():
x = symbol('x', '5 * 3 * float32')
assert summary(a=x.min(), b=x.max(), axis=0).dshape == \
dshape('3 * {a: float32, b: float32}')
assert summary(a=x.min(), b=x.max(), axis=1).dshape == \
dshape('5 * {a: float32, b: float32}')
assert summary(a=x.min(), b=x.max(), axis=1, keepdims=True).dshape == \
dshape('5 * 1 * {a: float32, b: float32}')
def test_summary_on_ndarray():
assert compute(summary(total=a.sum(), min=a.min()), ax) == \
(ax.min(), ax.sum())
result = compute(summary(total=a.sum(), min=a.min(), keepdims=True), ax)
expected = np.array([(ax.min(), ax.sum())],
dtype=[('min', 'float32'), ('total', 'float64')])
assert result.ndim == ax.ndim
assert eq(expected, result)
def test_summary_on_ndarray():
assert compute(summary(total=a.sum(), min=a.min()), ax) == \
(ax.min(), ax.sum())
result = compute(summary(total=a.sum(), min=a.min(), keepdims=True), ax)
expected = np.array([(ax.min(), ax.sum())],
dtype=[('min', 'float32'), ('total', 'float64')])
assert result.ndim == ax.ndim
assert eq(expected, result)
def test_summary_by():
expr = by(t.name, summary(count=t.id.count(), sum=t.amount.sum()))
assert str(compute(expr, df)) == \
str(DataFrame([['Alice', 2, 150],
['Bob', 1, 200]], columns=['name', 'count', 'sum']))
expr = by(t.name, summary(count=t.id.count(), sum=(t.amount + 1).sum()))
assert str(compute(expr, df)) == \
str(DataFrame([['Alice', 2, 152],
['Bob', 1, 201]], columns=['name', 'count', 'sum']))
def test_summary_by():
expr = by(t.name, summary(count=t.id.count(), sum=t.amount.sum()))
result = compute(expr, df)
expected = DataFrame([['Alice', 2, 150],
['Bob', 1, 200]], columns=['name', 'count', 'sum'])
expr = by(t.name, summary(count=t.id.count(), sum=(t.amount + 1).sum()))
result = compute(expr, df)
expected = DataFrame([['Alice', 2, 152],
['Bob', 1, 201]], columns=['name', 'count', 'sum'])
tm.assert_frame_equal(result, expected)
def test_summary_by():
expr = by(t.name, summary(count=t.id.count(), sum=t.amount.sum()))
result = compute(expr, df)
expected = DataFrame([['Alice', 2, 150], ['Bob', 1, 200]],
columns=['name', 'count', 'sum'])
expr = by(t.name, summary(count=t.id.count(), sum=(t.amount + 1).sum()))
result = compute(expr, df)
expected = DataFrame([['Alice', 2, 152], ['Bob', 1, 201]],
columns=['name', 'count', 'sum'])
tm.assert_frame_equal(result, expected)
def test_summary_by():
expr = by(t.name, summary(count=t.id.count(), sum=t.amount.sum()))
assert set(compute(expr, data)) == set([('Alice', 2, 150),
('Bob', 1, 200)])
expr = by(t.name, summary(count=t.id.count(), sum=(t.amount + 1).sum()))
assert set(compute(expr, data)) == set([('Alice', 2, 152),
('Bob', 1, 201)])
expr = by(t.name, summary(count=t.id.count(), sum=t.amount.sum() + 1))
assert set(compute(expr, data)) == set([('Alice', 2, 151),
('Bob', 1, 201)])
def test_summary_by():
expr = by(t.name, summary(count=t.id.count(), sum=t.amount.sum()))
assert set(compute(expr, data)) == set([('Alice', 2, 150),
('Bob', 1, 200)])
expr = by(t.name, summary(count=t.id.count(), sum=(t.amount + 1).sum()))
assert set(compute(expr, data)) == set([('Alice', 2, 152),
('Bob', 1, 201)])
expr = by(t.name, summary(count=t.id.count(), sum=t.amount.sum() + 1))
assert set(compute(expr, data)) == set([('Alice', 2, 151),
('Bob', 1, 201)])
def test_summary():
t = TableSymbol("t", "{id: int32, name: string, amount: int32}")
s = summary(total=t.amount.sum(), num=t.id.count())
assert s.dshape == dshape("{num: int32, total: int64}")
assert hash(s)
assert eval(str(s)).isidentical(s)
assert "summary(" in str(s)
assert "total=" in str(s)
assert "num=" in str(s)
assert str(t.amount.sum()) in str(s)
assert not summary(total=t.amount.sum())._child.isidentical(t.amount.sum())
assert iscollection(summary(total=t.amount.sum() + 1)._child.dshape)
def test_summary():
t = symbol('t', 'var * {id: int32, name: string, amount: int32}')
s = summary(total=t.amount.sum(), num=t.id.count())
assert s.dshape == dshape('{num: int32, total: int64}')
assert hash(s)
assert eval(str(s)).isidentical(s)
assert 'summary(' in str(s)
assert 'total=' in str(s)
assert 'num=' in str(s)
assert str(t.amount.sum()) in str(s)
assert not summary(total=t.amount.sum())._child.isidentical(t.amount.sum())
assert iscollection(summary(total=t.amount.sum() + 1)._child.dshape)
def test_summary_with_mean():
(chunk, chunk_expr), (agg, agg_expr) = split(t, summary(a=t.amount.count(),
b=t.id.mean() + 1))
assert chunk.schema == t.schema
assert chunk_expr.isidentical(summary(a=chunk.amount.count(),
b_total=chunk.id.sum(),
b_count=chunk.id.count(),
keepdims=True))
# assert not agg.schema == dshape('{a: int32, b: int32}')
expected = summary(a=agg.a.sum(),
b=(agg.b_total.sum() / agg.b_count.sum()) + 1)
assert agg_expr.isidentical(expected)
def test_summary():
t = TableSymbol('t', '{id: int32, name: string, amount: int32}')
s = summary(total=t.amount.sum(), num=t.id.count())
assert s.dshape == dshape('{num: int32, total: int32}')
assert hash(s)
assert eval(str(s)).isidentical(s)
assert 'summary(' in str(s)
assert 'total=' in str(s)
assert 'num=' in str(s)
assert str(t.amount.sum()) in str(s)
assert not summary(total=t.amount.sum()).child.isidentical(
t.amount.sum())
assert isinstance(summary(total=t.amount.sum() + 1).child, TableExpr)
def test_summary_with_mean():
(chunk, chunk_expr), (agg, agg_expr) = split(
t, summary(a=t.amount.count(), b=t.id.mean() + 1))
assert chunk.schema == t.schema
assert chunk_expr.isidentical(
summary(a=chunk.amount.count(),
b_total=chunk.id.sum(),
b_count=chunk.id.count(),
keepdims=True))
# assert not agg.schema == dshape('{a: int32, b: int32}')
expected = summary(a=agg.a.sum(),
b=(agg.b_total.sum() / agg.b_count.sum()) + 1)
assert agg_expr.isidentical(expected)
def test_elemwise_with_multiple_paths():
s = symbol('s', 'var * {x: int, y: int, z: int}')
expr = s.x.sum() / s.y.sum()
(chunk, chunk_expr), (agg, agg_expr) = split(s, expr)
assert chunk_expr.isidentical(summary(x=chunk.x.sum(), y=chunk.y.sum()))
assert agg_expr.isidentical(agg.x / agg.y)
def test_elemwise_with_multiple_paths():
s = symbol('s', 'var * {x: int, y: int, z: int}')
expr = s.x.sum() / s.y.sum()
(chunk, chunk_expr), (agg, agg_expr) = split(s, expr)
assert chunk_expr.isidentical(summary(x=chunk.x.sum(), y=chunk.y.sum()))
assert agg_expr.isidentical(agg.x / agg.y)
def test_lean_by_with_summary():
assert lean_projection(by(t.x, total=t.y.sum()))._child.isidentical(t[["x", "y"]])
tt = t[["x", "y"]]
result = lean_projection(by(t.x, a=t.y.sum(), b=t.z.sum())[["x", "a"]])
expected = Projection(By(Field(tt, "x"), summary(a=sum(Field(tt, "y")))), ("x", "a"))
assert result.isidentical(expected)
def test_summary_clean():
t2 = t[t.amount > 0]
expr = summary(a=t2.amount.sum(), b=t2.id.count())
result = str(compute(expr, s))
assert normalize(result) == normalize("""
SELECT sum(accounts.amount) as a, count(accounts.id) as b
FROM accounts
WHERE accounts.amount > :amount_1""")
def test_summary_by():
expr = by(t.name, summary(a=t.amount.sum(), b=t.id.count()))
result = str(compute(expr, s))
assert 'sum(accounts.amount) as a' in result.lower()
assert 'count(accounts.id) as b' in result.lower()
assert 'group by accounts.name' in result.lower()
def test_summary_clean():
t2 = t[t.amount > 0]
expr = summary(a=t2.amount.sum(), b=t2.id.count())
result = str(compute(expr, s))
assert normalize(result) == normalize("""
SELECT sum(accounts.amount) as a, count(accounts.id) as b
FROM accounts
WHERE accounts.amount > :amount_1""")
def test_summary_by():
expr = by(t.name, summary(a=t.amount.sum(), b=t.id.count()))
result = str(compute(expr, s))
assert 'sum(accounts.amount) as a' in result.lower()
assert 'count(accounts.id) as b' in result.lower()
assert 'group by accounts.name' in result.lower()
def test_mean():
(chunk, chunk_expr), (agg, agg_expr) = split(t, t.amount.mean())
assert chunk.schema == t.schema
assert chunk_expr.isidentical(
summary(total=chunk.amount.sum(),
count=chunk.amount.count(),
keepdims=True))
assert isrecord(agg.dshape.measure)
assert agg_expr.isidentical(agg.total.sum() / agg['count'].sum())
def test_complex_summaries():
t = symbol('t', '100 * {a: int, b: int}')
(chunk, chunk_expr), (agg, agg_expr) = split(t, summary(q=t.a.mean(),
w=t.a.std(),
e=t.a.sum()))
assert chunk_expr.isidentical(summary(e=chunk.a.sum(),
q_count=chunk.a.count(),
q_total=chunk.a.sum(),
w_n=chunk.a.count(),
w_x=chunk.a.sum(),
w_x2=(chunk.a ** 2).sum(),
keepdims=True))
expected = summary(e=agg.e.sum(),
q=agg.q_total.sum() / agg.q_count.sum(),
w=sqrt((agg.w_x2.sum() / agg.w_n.sum())
- (agg.w_x.sum() / agg.w_n.sum()) ** 2))
assert agg_expr.isidentical(expected)
def test_complex_summaries():
t = symbol('t', '100 * {a: int, b: int}')
(chunk, chunk_expr), (agg, agg_expr) = split(
t, summary(q=t.a.mean(), w=t.a.std(), e=t.a.sum()))
assert chunk_expr.isidentical(
summary(e=chunk.a.sum(),
q_count=chunk.a.count(),
q_total=chunk.a.sum(),
w_n=chunk.a.count(),
w_x=chunk.a.sum(),
w_x2=(chunk.a**2).sum(),
keepdims=True))
expected = summary(e=agg.e.sum(),
q=agg.q_total.sum() / agg.q_count.sum(),
w=sqrt((agg.w_x2.sum() / agg.w_n.sum()) -
(agg.w_x.sum() / agg.w_n.sum())**2))
assert agg_expr.isidentical(expected)
def test_mean():
(chunk, chunk_expr), (agg, agg_expr) = split(t, t.amount.mean())
assert chunk.schema == t.schema
assert chunk_expr.isidentical(summary(total=chunk.amount.sum(),
count=chunk.amount.count(),
keepdims=True))
assert isrecord(agg.dshape.measure)
assert agg_expr.isidentical(agg.total.sum() / agg.count.sum())
def test_lean_by_with_summary():
assert lean_projection(by(t.x, total=t.y.sum()))._child.isidentical(
t[['x', 'y']], )
tt = t[['x', 'y']]
result = lean_projection(by(t.x, a=t.y.sum(), b=t.z.sum())[['x', 'a']])
expected = Projection(
By(Field(tt, 'x'), summary(a=sum(Field(tt, 'y')))),
('x', 'a'),
)
assert result.isidentical(expected)
def test_summary_on_ndarray_with_axis():
for axis in [0, 1, (1, 0)]:
expr = summary(total=a.sum(), min=a.min(), axis=axis)
result = compute(expr, ax)
shape, dtype = to_numpy(expr.dshape)
expected = np.empty(shape=shape, dtype=dtype)
expected['total'] = ax.sum(axis=axis)
expected['min'] = ax.min(axis=axis)
assert eq(result, expected)
def test_summary_on_ndarray_with_axis():
for axis in [0, 1, (1, 0)]:
expr = summary(total=a.sum(), min=a.min(), axis=axis)
result = compute(expr, ax)
shape, dtype = to_numpy(expr.dshape)
expected = np.empty(shape=shape, dtype=dtype)
expected['total'] = ax.sum(axis=axis)
expected['min'] = ax.min(axis=axis)
assert eq(result, expected)
def test_std():
(chunk, chunk_expr), (agg, agg_expr) = split(t, t.amount.std())
assert chunk.schema == t.schema
assert chunk_expr.isidentical(summary(x=chunk.amount.sum(),
x2=(chunk.amount ** 2).sum(),
n=chunk.amount.count(),
keepdims=True))
assert isrecord(agg.dshape.measure)
assert agg_expr.isidentical(sqrt((agg.x2.sum() / (agg.n.sum())
- (agg.x.sum() / (agg.n.sum())) ** 2)))
def test_var():
(chunk, chunk_expr), (agg, agg_expr) = split(t, t.amount.var())
assert chunk.schema == t.schema
assert chunk_expr.isidentical(
summary(x=chunk.amount.sum(),
x2=(chunk.amount**2).sum(),
n=chunk.amount.count(),
keepdims=True))
assert isrecord(agg.dshape.measure)
assert agg_expr.isidentical(
(agg.x2.sum() / (agg.n.sum()) - (agg.x.sum() / (agg.n.sum()))**2))
def test_path_split():
expr = t.amount.sum() + 1
assert path_split(t, expr).isidentical(t.amount.sum())
expr = t.amount.distinct().sort()
assert path_split(t, expr).isidentical(t.amount.distinct())
t2 = transform(t, id=t.id * 2)
expr = by(t2.id, amount=t2.amount.sum()).amount + 1
assert path_split(t, expr).isidentical(by(t2.id, amount=t2.amount.sum()))
expr = count(t.amount.distinct())
assert path_split(t, expr).isidentical(t.amount.distinct())
expr = summary(total=t.amount.sum())
assert path_split(t, expr).isidentical(expr)
def test_path_split():
expr = t.amount.sum() + 1
assert path_split(t, expr).isidentical(t.amount.sum())
expr = t.amount.distinct().sort()
assert path_split(t, expr).isidentical(t.amount.distinct())
t2 = transform(t, id=t.id * 2)
expr = by(t2.id, amount=t2.amount.sum()).amount + 1
assert path_split(t, expr).isidentical(by(t2.id, amount=t2.amount.sum()))
expr = count(t.amount.distinct())
assert path_split(t, expr).isidentical(t.amount.distinct())
expr = summary(total=t.amount.sum())
assert path_split(t, expr).isidentical(expr)
def test_join_on_same_table():
metadata = sa.MetaData()
T = sa.Table(
'tab',
metadata,
sa.Column('a', sa.Integer),
sa.Column('b', sa.Integer),
)
t = symbol('tab', discover(T))
expr = join(t, t, 'a')
result = compute(expr, {t: T})
assert normalize(str(result)) == normalize("""
SELECT tab_left.a, tab_left.b, tab_right.b
FROM tab AS tab_left JOIN tab AS tab_right
ON tab_left.a = tab_right.a
""")
expr = join(t, t, 'a').b_left.sum()
result = compute(expr, {t: T})
assert normalize(str(result)) == normalize("""
with alias as
(select tab_left.b as b
from tab as tab_left
join tab as tab_right
on tab_left.a = tab_right.a)
select sum(alias.b) as b_left_sum from alias""")
expr = join(t, t, 'a')
expr = summary(total=expr.a.sum(), smallest=expr.b_right.min())
result = compute(expr, {t: T})
assert normalize(str(result)) == normalize("""
SELECT min(tab_right.b) as smallest, sum(tab_left.a) as total
FROM tab AS tab_left JOIN tab AS tab_right
ON tab_left.a = tab_right.a
""")
def test_join_on_same_table():
metadata = sa.MetaData()
T = sa.Table('tab', metadata,
sa.Column('a', sa.Integer),
sa.Column('b', sa.Integer),
)
t = symbol('tab', discover(T))
expr = join(t, t, 'a')
result = compute(expr, {t: T})
assert normalize(str(result)) == normalize("""
SELECT tab_left.a, tab_left.b, tab_right.b
FROM tab AS tab_left JOIN tab AS tab_right
ON tab_left.a = tab_right.a
""")
expr = join(t, t, 'a').b_left.sum()
result = compute(expr, {t: T})
assert normalize(str(result)) == normalize("""
with alias as
(select tab_left.b as b
from tab as tab_left
join tab as tab_right
on tab_left.a = tab_right.a)
select sum(alias.b) as b_left_sum from alias""")
expr = join(t, t, 'a')
expr = summary(total=expr.a.sum(), smallest=expr.b_right.min())
result = compute(expr, {t: T})
assert normalize(str(result)) == normalize("""
SELECT min(tab_right.b) as smallest, sum(tab_left.a) as total
FROM tab AS tab_left JOIN tab AS tab_right
ON tab_left.a = tab_right.a
""")
def test_summary():
(chunk, chunk_expr), (agg, agg_expr) = split(
t, summary(a=t.amount.count(), b=t.id.sum() + 1))
assert chunk.schema == t.schema
assert chunk_expr.isidentical(
summary(a=chunk.amount.count(), b=chunk.id.sum(), keepdims=True))
# assert not agg.schema == dshape('{a: int32, b: int32}')
assert agg_expr.isidentical(summary(a=agg.a.sum(), b=agg.b.sum() + 1))
(chunk, chunk_expr), (agg, agg_expr) = \
split(t, summary(total=t.amount.sum()))
assert chunk_expr.isidentical(
summary(total=chunk.amount.sum(), keepdims=True))
assert agg_expr.isidentical(summary(total=agg.total.sum()))
def test_summary():
(chunk, chunk_expr), (agg, agg_expr) = split(t, summary(a=t.amount.count(),
b=t.id.sum() + 1))
assert chunk.schema == t.schema
assert chunk_expr.isidentical(summary(a=chunk.amount.count(),
b=chunk.id.sum(), keepdims=True))
# assert not agg.schema == dshape('{a: int32, b: int32}')
assert agg_expr.isidentical(summary(a=agg.a.sum(),
b=agg.b.sum() + 1))
(chunk, chunk_expr), (agg, agg_expr) = \
split(t, summary(total=t.amount.sum()))
assert chunk_expr.isidentical(summary(total=chunk.amount.sum(),
keepdims=True))
assert agg_expr.isidentical(summary(total=agg.total.sum()))
def test_summary_keepdims():
x = symbol('x', '5 * 3 * float32')
assert summary(a=x.min(), b=x.max()).dshape == \
dshape('{a: float32, b: float32}')
assert summary(a=x.min(), b=x.max(), keepdims=True).dshape == \
dshape('1 * 1 * {a: float32, b: float32}')
def test_summary_with_multiple_children():
t = symbol('t', 'var * {x: int, y: int, z: int}')
assert summary(a=t.x.sum() + t.y.sum())._child.isidentical(t)
def test_summary():
expr = summary(count=t.id.count(), sum=t.amount.sum())
assert str(compute(expr, df)) == str(Series({'count': 3, 'sum': 350}))
def test_by_summary():
t = symbol('t', 'var * {name: string, amount: int32, id: int32}')
a = by(t['name'], sum=sum(t['amount']))
b = by(t['name'], summary(sum=sum(t['amount'])))
assert a.isidentical(b)
def test_summary():
expr = summary(count=t.id.count(), sum=t.amount.sum())
assert_series_equal(compute(expr, df), Series({'count': 3, 'sum': 350}))
def test_summary():
expr = summary(count=t.id.count(), sum=t.amount.sum())
assert_series_equal(compute(expr, df), Series({'count': 3, 'sum': 350}))
def test_summary_keepdims():
expr = summary(count=t.id.count(), sum=t.amount.sum(), keepdims=True)
expected = DataFrame([[3, 350]], columns=['count', 'sum'])
tm.assert_frame_equal(compute(expr, df), expected)
def test_summary():
expr = summary(count=t.id.count(), sum=t.amount.sum())
assert compute(expr, data) == (3, 350)
def test_summary_by_reduction_arithmetic():
expr = by(t.name, summary(count=t.id.count(), sum=t.amount.sum() + 1))
assert str(compute(expr, df)) == \
str(DataFrame([['Alice', 2, 151],
['Bob', 1, 202]], columns=['name', 'count', 'sum']))
def test_summary_by_reduction_arithmetic():
expr = by(t.name, summary(count=t.id.count(), sum=t.amount.sum() + 1))
result = compute(expr, df)
expected = DataFrame([['Alice', 2, 151],
['Bob', 1, 201]], columns=['name', 'count', 'sum'])
tm.assert_frame_equal(result, expected)
def test_summary():
expr = summary(count=t.id.count(), sum=t.amount.sum())
eq(compute(expr, ddf), pd.Series({'count': 3, 'sum': 350}))
def test_summary_keepdims():
expr = summary(count=t.id.count(), sum=t.amount.sum(), keepdims=True)
expected = DataFrame([[3, 350]], columns=['count', 'sum'])
tm.assert_frame_equal(compute(expr, df), expected)
def test_summary_by_reduction_arithmetic():
expr = by(t.name, summary(count=t.id.count(), sum=t.amount.sum() + 1))
result = compute(expr, df)
expected = DataFrame([['Alice', 2, 151], ['Bob', 1, 201]],
columns=['name', 'count', 'sum'])
tm.assert_frame_equal(result, expected)
def test_by_summary():
t = TableSymbol('t', '{name: string, amount: int32, id: int32}')
a = by(t['name'], sum=sum(t['amount']))
b = by(t['name'], summary(sum=sum(t['amount'])))
assert a.isidentical(b)
def test_summary_str():
x = symbol('x', '5 * 3 * float32')
assert 'keepdims' not in str(summary(a=x.min(), b=x.max()))
