NumPy 1.24 Release Notes — NumPy v2.0 Manual (2024)

Deprecate fastCopyAndTranspose and PyArray_CopyAndTranspose#

The numpy.fastCopyAndTranspose function has been deprecated. Use thecorresponding copy and transpose methods directly:

arr.T.copy()

The underlying C function PyArray_CopyAndTranspose has also been deprecatedfrom the NumPy C-API.

(gh-22313)

Conversion of out-of-bound Python integers#

Attempting a conversion from a Python integer to a NumPy value will now alwayscheck whether the result can be represented by NumPy. This means the followingexamples will fail in the future and give a DeprecationWarning now:

np.uint8(-1)np.array([3000], dtype=np.int8)

Many of these did succeed before. Such code was mainly useful for unsignedintegers with negative values such as np.uint8(-1) givingnp.iinfo(np.uint8).max.

Note that conversion between NumPy integers is unaffected, so thatnp.array(-1).astype(np.uint8) continues to work and use C integer overflowlogic. For negative values, it will also work to view the array:np.array(-1, dtype=np.int8).view(np.uint8).In some cases, using np.iinfo(np.uint8).max or val % 2**8 may alsowork well.

In rare cases input data may mix both negative values and very large unsignedvalues (i.e. -1 and 2**63). There it is unfortunately necessaryto use % on the Python value or use signed or unsigned conversiondepending on whether negative values are expected.

(gh-22385)

Deprecate msort#

The numpy.msort function is deprecated. Use np.sort(a, axis=0) instead.

(gh-22456)

np.str0 and similar are now deprecated#

The scalar type aliases ending in a 0 bit size: np.object0, np.str0,np.bytes0, np.void0, np.int0, np.uint0 as well as np.bool8are now deprecated and will eventually be removed.

(gh-22607)

array.fill(scalar) may behave slightly different#

numpy.ndarray.fill may in some cases behave slightly different now due tothe fact that the logic is aligned with item assignment:

arr = np.array([1]) # with any dtype/valuearr.fill(scalar)# is now identical to:arr[0] = scalar

Previously casting may have produced slightly different answers when usingvalues that could not be represented in the target dtype or when the targethad object dtype.

(gh-20924)

Subarray to object cast now copies#

Casting a dtype that includes a subarray to an object will now ensure a copy ofthe subarray. Previously an unsafe view was returned:

arr = np.ones(3, dtype=[("f", "i", 3)])subarray_fields = arr.astype(object)[0]subarray = subarray_fields[0] # "f" fieldnp.may_share_memory(subarray, arr)

Is now always false. While previously it was true for the specific cast.

(gh-21925)

Returned arrays respect uniqueness of dtype kwarg objects#

When the dtype keyword argument is used with array orasarray, the dtype of the returned array now always exactlymatches the dtype provided by the caller.

In some cases this change means that a view rather than the input array isreturned. The following is an example for this on 64bit Linux where longand longlong are the same precision but different dtypes:

>>> arr = np.array([1, 2, 3], dtype="long")>>> new_dtype = np.dtype("longlong")>>> new = np.asarray(arr, dtype=new_dtype)>>> new.dtype is new_dtypeTrue>>> new is arrFalse

Before the change, the dtype did not match because new is arr wasTrue.

(gh-21995)

DLPack export raises BufferError#

When an array buffer cannot be exported via DLPack a BufferError is nowalways raised where previously TypeError or RuntimeError was raised.This allows falling back to the buffer protocol or __array_interface__ whenDLPack was tried first.

(gh-22542)

NumPy builds are no longer tested on GCC-6#

Ubuntu 18.04 is deprecated for GitHub actions and GCC-6 is not available onUbuntu 20.04, so builds using that compiler are no longer tested. We still testbuilds using GCC-7 and GCC-8.

(gh-22598)

New attribute symbol added to polynomial classes#

The polynomial classes in the numpy.polynomial package have a newsymbol attribute which is used to represent the indeterminate of thepolynomial. This can be used to change the value of the variable whenprinting:

>>> P_y = np.polynomial.Polynomial([1, 0, -1], symbol="y")>>> print(P_y)1.0 + 0.0·y¹ - 1.0·y²

Note that the polynomial classes only support 1D polynomials, so operationsthat involve polynomials with different symbols are disallowed when the resultwould be multivariate:

>>> P = np.polynomial.Polynomial([1, -1]) # default symbol is "x">>> P_z = np.polynomial.Polynomial([1, 1], symbol="z")>>> P * P_zTraceback (most recent call last) ...ValueError: Polynomial symbols differ

The symbol can be any valid Python identifier. The default is symbol=x,consistent with existing behavior.

(gh-16154)

F2PY support for Fortran character strings#

F2PY now supports wrapping Fortran functions with:

• character (e.g. character x)

• character array (e.g. character, dimension(n) :: x)

• character string (e.g. character(len=10) x)

• and character string array (e.g. character(len=10), dimension(n, m) :: x)

arguments, including passing Python unicode strings as Fortran character stringarguments.

(gh-19388)

New function np.show_runtime#

A new function numpy.show_runtime has been added to display the runtimeinformation of the machine in addition to numpy.show_config which displaysthe build-related information.

(gh-21468)

strict option for testing.assert_array_equal#

The strict option is now available for testing.assert_array_equal.Setting strict=True will disable the broadcasting behaviour for scalars andensure that input arrays have the same data type.

(gh-21595)

New parameter equal_nan added to np.unique#

np.unique was changed in 1.21 to treat all NaN values as equal andreturn a single NaN. Setting equal_nan=False will restore pre-1.21behavior to treat NaNs as unique. Defaults to True.

(gh-21623)

casting and dtype keyword arguments for numpy.stack#

The casting and dtype keyword arguments are now available fornumpy.stack. To use them, write np.stack(..., dtype=None,casting='same_kind').

casting and dtype keyword arguments for numpy.vstack#

The casting and dtype keyword arguments are now available fornumpy.vstack. To use them, write np.vstack(..., dtype=None,casting='same_kind').

casting and dtype keyword arguments for numpy.hstack#

The casting and dtype keyword arguments are now available fornumpy.hstack. To use them, write np.hstack(..., dtype=None,casting='same_kind').

(gh-21627)

The bit generator underlying the singleton RandomState can be changed#

The singleton RandomState instance exposed in the numpy.random moduleis initialized at startup with the MT19937 bit generator. The new functionset_bit_generator allows the default bit generator to be replaced with auser-provided bit generator. This function has been introduced to provide amethod allowing seamless integration of a high-quality, modern bit generator innew code with existing code that makes use of the singleton-provided randomvariate generating functions. The companion function get_bit_generatorreturns the current bit generator being used by the singleton RandomState.This is provided to simplify restoring the original source of randomness ifrequired.

The preferred method to generate reproducible random numbers is to use a modernbit generator in an instance of Generator. The function default_rngsimplifies instantiation:

>>> rg = np.random.default_rng(3728973198)>>> rg.random()

The same bit generator can then be shared with the singleton instance so thatcalling functions in the random module will use the same bit generator:

>>> orig_bit_gen = np.random.get_bit_generator()>>> np.random.set_bit_generator(rg.bit_generator)>>> np.random.normal()

The swap is permanent (until reversed) and so any call to functions in therandom module will use the new bit generator. The original can be restoredif required for code to run correctly:

>>> np.random.set_bit_generator(orig_bit_gen)

(gh-21976)

np.void now has a dtype argument#

NumPy now allows constructing structured void scalars directly bypassing the dtype argument to np.void.

(gh-22316)

F2PY Improvements#

• The generated extension modules don’t use the deprecated NumPy-C API anymore

• Improved f2py generated exception messages

• Numerous bug and flake8 warning fixes

• various CPP macros that one can use within C-expressions of signature filesare prefixed with f2py_. For example, one should use f2py_len(x)instead of len(x)

• A new construct character(f2py_len=...) is introduced to supportreturning assumed length character strings (e.g. character(len=*)) fromwrapper functions

A hook to support rewriting f2py internal data structures after reading allits input files is introduced. This is required, for instance, for BC of SciPysupport where character arguments are treated as character strings arguments inC expressions.

(gh-19388)

IBM zSystems Vector Extension Facility (SIMD)#

Added support for SIMD extensions of zSystem (z13, z14, z15), through theuniversal intrinsics interface. This support leads to performance improvementsfor all SIMD kernels implemented using the universal intrinsics, including thefollowing operations: rint, floor, trunc, ceil, sqrt, absolute, square,reciprocal, tanh, sin, cos, equal, not_equal, greater, greater_equal, less,less_equal, maximum, minimum, fmax, fmin, argmax, argmin, add, subtract,multiply, divide.

(gh-20913)

NumPy now gives floating point errors in casts#

In most cases, NumPy previously did not give floating point warnings or errorswhen these happened during casts. For examples, casts like:

np.array([2e300]).astype(np.float32) # overflow for float32np.array([np.inf]).astype(np.int64)

Should now generally give floating point warnings. These warnings should warnthat floating point overflow occurred. For errors when converting floatingpoint values to integers users should expect invalid value warnings.

Users can modify the behavior of these warnings using np.errstate.

Note that for float to int casts, the exact warnings that are given maybe platform dependent. For example:

arr = np.full(100, fill_value=1000, dtype=np.float64)arr.astype(np.int8)

May give a result equivalent to (the intermediate cast means no warning isgiven):

arr.astype(np.int64).astype(np.int8)

May return an undefined result, with a warning set:

RuntimeWarning: invalid value encountered in cast

The precise behavior is subject to the C99 standard and its implementation inboth software and hardware.

(gh-21437)

F2PY supports the value attribute#

The Fortran standard requires that variables declared with the valueattribute must be passed by value instead of reference. F2PY now supports thisuse pattern correctly. So integer, intent(in), value :: x in Fortran codeswill have correct wrappers generated.

(gh-21807)

Added pickle support for third-party BitGenerators#

The pickle format for bit generators was extended to allow each bit generatorto supply its own constructor when during pickling. Previous versions of NumPyonly supported unpickling Generator instances created with one of the coreset of bit generators supplied with NumPy. Attempting to unpickle aGenerator that used a third-party bit generators would fail since theconstructor used during the unpickling was only aware of the bit generatorsincluded in NumPy.

(gh-22014)

arange() now explicitly fails with dtype=str#

Previously, the np.arange(n, dtype=str) function worked for n=1 andn=2, but would raise a non-specific exception message for other values ofn. Now, it raises a TypeError informing that arange does not supportstring dtypes:

>>> np.arange(2, dtype=str)Traceback (most recent call last) ...TypeError: arange() not supported for inputs with DType <class 'numpy.dtype[str_]'>.

(gh-22055)

numpy.typing protocols are now runtime checkable#

The protocols used in numpy.typing.ArrayLike and numpy.typing.DTypeLikeare now properly marked as runtime checkable, making them easier to use forruntime type checkers.

(gh-22357)

Faster version of np.isin and np.in1d for integer arrays#

np.in1d (used by np.isin) can now switch to a faster algorithm (up to>10x faster) when it is passed two integer arrays. This is often automaticallyused, but you can use kind="sort" or kind="table" to force the old ornew method, respectively.

(gh-12065)

Faster comparison operators#

The comparison functions (numpy.equal, numpy.not_equal, numpy.less,numpy.less_equal, numpy.greater and numpy.greater_equal) are nowmuch faster as they are now vectorized with universal intrinsics. For a CPUwith SIMD extension AVX512BW, the performance gain is up to 2.57x, 1.65x and19.15x for integer, float and boolean data types, respectively (with N=50000).

(gh-21483)

Better reporting of integer division overflow#

Integer division overflow of scalars and arrays used to provide aRuntimeWarning and the return value was undefined leading to crashes atrare occasions:

>>> np.array([np.iinfo(np.int32).min]*10, dtype=np.int32) // np.int32(-1)<stdin>:1: RuntimeWarning: divide by zero encountered in floor_dividearray([0, 0, 0, 0, 0, 0, 0, 0, 0, 0], dtype=int32)

Integer division overflow now returns the input dtype’s minimum value and raisethe following RuntimeWarning:

>>> np.array([np.iinfo(np.int32).min]*10, dtype=np.int32) // np.int32(-1)<stdin>:1: RuntimeWarning: overflow encountered in floor_dividearray([-2147483648, -2147483648, -2147483648, -2147483648, -2147483648, -2147483648, -2147483648, -2147483648, -2147483648, -2147483648], dtype=int32)

(gh-21506)

masked_invalid now modifies the mask in-place#

When used with copy=False, numpy.ma.masked_invalid now modifies theinput masked array in-place. This makes it behave identically tomasked_where and better matches the documentation.

(gh-22046)

nditer/NpyIter allows all allocating all operands#

The NumPy iterator available through np.nditer in Python and as NpyIterin C now supports allocating all arrays. The iterator shape defaults to ()in this case. The operands dtype must be provided, since a “common dtype”cannot be inferred from the other inputs.

(gh-22457)