Skip to content

bpo-40889: special-case dict.items() ^ dict.items() for performance #20718

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
methane merged 10 commits into from
Jun 10, 2020
Merged

bpo-40889: special-case dict.items() ^ dict.items() for performance #20718

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
9d7406f
special-case dict.items() ^ dict.items() for performance
sweeneyde Jun 8, 2020
9dca79c
Add tests for dict_items.__xor__
sweeneyde Jun 8, 2020
eb4c1cd
📜🤖 Added by blurb_it.
blurb-it[bot] Jun 8, 2020
537575d
Be subclass-friendly
sweeneyde Jun 9, 2020
b10c78f
Merge branch 'dictitems_xor' of https://github.com/sweeneyde/cpython …
sweeneyde Jun 9, 2020
60aa7f6
Use PyTuple_Pack
sweeneyde Jun 9, 2020
0036167
Change: XDECREF --> DECREF
sweeneyde Jun 9, 2020
1ef9f93
Iterator protocol --> _PyDict_Next
sweeneyde Jun 9, 2020
fa0426a
Revert "Iterator protocol --> _PyDict_Next"
sweeneyde Jun 9, 2020
30e3ae2
Use the items() iterator rather than PyDict_Items()
sweeneyde Jun 9, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 10 additions & 0 deletions Lib/test/test_dict.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -697,6 +697,16 @@ def test_dictview_set_operations_on_items(self):
self.assertEqual(k1 ^ k2, {(3,3)})
self.assertEqual(k1 ^ k3, {(1,1), (2,2), (4,4)})

def test_items_symmetric_difference(self):
rr = random.randrange
for _ in range(100):
left = {x:rr(3) for x in range(20) if rr(2)}
right = {x:rr(3) for x in range(20) if rr(2)}
with self.subTest(left=left, right=right):
expected = set(left.items()) ^ set(right.items())
actual = left.items() ^ right.items()
self.assertEqual(actual, expected)

def test_dictview_mixed_set_operations(self):
# Just a few for .keys()
self.assertTrue({1:1}.keys() == {1})
Expand Down
1 change: 1 addition & 0 deletions Misc/NEWS.d/next/Core and Builtins/2020-06-08-22-46-33.bpo-40889.vIBl-W.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1 @@
Improved the performance of symmetric difference operations on dictionary item views. Patch by Dennis Sweeney.
90 changes: 90 additions & 0 deletions Objects/dictobject.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4409,9 +4409,99 @@ dictviews_or(PyObject* self, PyObject *other)
return result;
}

static PyObject *
dictitems_xor(PyObject *self, PyObject *other)
{
assert(PyDictItems_Check(self));
assert(PyDictItems_Check(other));
PyObject *d1 = (PyObject *)((_PyDictViewObject *)self)->dv_dict;
PyObject *d2 = (PyObject *)((_PyDictViewObject *)other)->dv_dict;

PyObject *temp_dict = PyDict_Copy(d1);
if (temp_dict == NULL) {
return NULL;
}
PyObject *result_set = PySet_New(NULL);
if (result_set == NULL) {
Py_CLEAR(temp_dict);
return NULL;
}

PyObject *key = NULL, *val1 = NULL, *val2 = NULL;

@rhettinger rhettinger Jun 10, 2020

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I generally expect these declarations to go at the top of the function (it's not required but it is the usual style in our C code).

Py_ssize_t pos = 0;
Py_hash_t hash;

while (_PyDict_Next(d2, &pos, &key, &val2, &hash)) {
Py_INCREF(key);
Py_INCREF(val2);
val1 = _PyDict_GetItem_KnownHash(temp_dict, key, hash);

int to_delete;
if (val1 == NULL) {
if (PyErr_Occurred()) {
goto error;
}
to_delete = 0;
}
else {
Py_INCREF(val1);
to_delete = PyObject_RichCompareBool(val1, val2, Py_EQ);
if (to_delete < 0) {
goto error;
}
}

if (to_delete) {
if (_PyDict_DelItem_KnownHash(temp_dict, key, hash) < 0) {
goto error;
}
}
else {
PyObject *pair = PyTuple_Pack(2, key, val2);
if (pair == NULL) {
goto error;
}
if (PySet_Add(result_set, pair) < 0) {
Py_DECREF(pair);
goto error;
}
Py_DECREF(pair);
}
Py_DECREF(key);
Py_XDECREF(val1);
Py_DECREF(val2);
}
key = val1 = val2 = NULL;

_Py_IDENTIFIER(items);
PyObject *remaining_pairs = _PyObject_CallMethodIdNoArgs(temp_dict,
&PyId_items);
if (remaining_pairs == NULL) {
goto error;
}
if (_PySet_Update(result_set, remaining_pairs) < 0) {
Py_DECREF(remaining_pairs);
goto error;
}
Py_DECREF(temp_dict);
Py_DECREF(remaining_pairs);
return result_set;

error:
Py_XDECREF(temp_dict);
Py_XDECREF(result_set);
Py_XDECREF(key);
Py_XDECREF(val1);
Py_XDECREF(val2);
return NULL;
}

static PyObject*
dictviews_xor(PyObject* self, PyObject *other)
{
if (PyDictItems_Check(self) && PyDictItems_Check(other)) {
return dictitems_xor(self, other);
}
PyObject *result = dictviews_to_set(self);
if (result == NULL) {
return NULL;
Expand Down