Skip to content

ENH: add fast path for str(scalar_int) #23746

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
charris merged 2 commits into from
May 13, 2023
Merged

ENH: add fast path for str(scalar_int) #23746

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
670842b
add fast path for str(scalar_int)
mattip May 10, 2023
5e983f2
BUG: typo, linting
mattip May 10, 2023
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
12 changes: 12 additions & 0 deletions benchmarks/benchmarks/bench_scalar.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -65,3 +65,15 @@ def time_add_other_and_int32arr(self, typename):
other + int32
other + int32
other + int32


class ScalarStr(Benchmark):
# Test scalar to str conversion
params = [TYPES1]
param_names = ["type"]

def setup(self, typename):
self.a = np.array([100] * 100, dtype=typename)

def time_str_repr(self, typename):
res = [str(x) for x in self.a]
38 changes: 37 additions & 1 deletion numpy/core/src/multiarray/scalartypes.c.src
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -279,7 +279,43 @@ static PyObject *
genint_type_str(PyObject *self)
{
PyObject *item, *item_str;
item = gentype_generic_method(self, NULL, NULL, "item");
PyArray_Descr *descr = PyArray_DescrFromTypeObject((PyObject *)Py_TYPE(self));
void *val = scalar_value(self, descr);
switch (descr->type_num) {
case NPY_BYTE:
item = PyLong_FromLong(*(int8_t *)val);
break;
case NPY_UBYTE:
item = PyLong_FromUnsignedLong(*(uint8_t *)val);
break;
case NPY_SHORT:
item = PyLong_FromLong(*(int16_t *)val);
break;
case NPY_USHORT:
item = PyLong_FromUnsignedLong(*(uint16_t *)val);
break;
case NPY_INT:
item = PyLong_FromLong(*(int32_t *)val);
break;
case NPY_UINT:
item = PyLong_FromUnsignedLong(*(uint32_t *)val);
break;
case NPY_LONG:
item = PyLong_FromLong(*(int64_t *)val);
break;
case NPY_ULONG:
item = PyLong_FromUnsignedLong(*(uint64_t *)val);
break;
case NPY_LONGLONG:
item = PyLong_FromLongLong(*(long long *)val);
break;
case NPY_ULONGLONG:
item = PyLong_FromUnsignedLongLong(*(unsigned long long *)val);
break;
default:
item = gentype_generic_method(self, NULL, NULL, "item");
break;
Copy link
Copy Markdown
Member Author

@mattip mattip May 10, 2023

Choose a reason for hiding this comment

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

A fallback for some as-yet-unknown user defined int dtype

Copy link
Copy Markdown
Member

@seberg seberg May 11, 2023

Choose a reason for hiding this comment

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

We need to generate a function for each type anyway. I don't think that str(self.item()) is a good default to begin with.

That said, I don't mind just doing this, I will just delete it again on main in the next 2 months hopefully. (The work should already be in the repr PR, and its time to push that after branching.)

}
if (item == NULL) {
return NULL;
}
Expand Down
5 changes: 3 additions & 2 deletions numpy/core/tests/test_arrayprint.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -258,8 +258,7 @@ def _format_function(x):
assert_(np.array2string(s, formatter={'numpystr':lambda s: s*2}) ==
'[abcabc defdef]')


def test_structure_format(self):
def test_structure_format_mixed(self):
Copy link
Copy Markdown
Member Author

@mattip mattip May 10, 2023

Choose a reason for hiding this comment

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

I split this test into 3, since the other test functions were not related to this one. I hit this by accident when I was trying out a different version of the code and printing was failing.

dt = np.dtype([('name', np.str_, 16), ('grades', np.float64, (2,))])
x = np.array([('Sarah', (8.0, 7.0)), ('John', (6.0, 7.0))], dtype=dt)
assert_equal(np.array2string(x),
Expand Down Expand Up @@ -301,6 +300,7 @@ def test_structure_format(self):
( 'NaT',) ( 'NaT',) ( 'NaT',)]""")
)

def test_structure_format_int(self):
# See #8160
struct_int = np.array([([1, -1],), ([123, 1],)], dtype=[('B', 'i4', 2)])
assert_equal(np.array2string(struct_int),
Expand All @@ -310,6 +310,7 @@ def test_structure_format(self):
assert_equal(np.array2string(struct_2dint),
"[([[ 0, 1], [ 2, 3]],) ([[12, 0], [ 0, 0]],)]")

def test_structure_format_float(self):
# See #8172
array_scalar = np.array(
(1., 2.1234567890123456789, 3.), dtype=('f8,f8,f8'))
Expand Down