numpysupported.rst

Supported NumPy features

One objective of Numba is having a seamless integration with NumPy. NumPy arrays provide an efficient storage method for homogeneous sets of data. NumPy dtypes provide type information useful when compiling, and the regular, structured storage of potentially large amounts of data in memory provides an ideal memory layout for code generation. Numba excels at generating code that executes on top of NumPy arrays.

NumPy support in Numba comes in many forms:

• Numba understands calls to NumPy ufuncs and is able to generate equivalent native code for many of them.
• NumPy arrays are directly supported in Numba. Access to Numpy arrays is very efficient, as indexing is lowered to direct memory accesses when possible.
• Numba is able to generate ufuncs and gufuncs. This means that it is possible to implement ufuncs and gufuncs within Python, getting speeds comparable to that of ufuncs/gufuncs implemented in C extension modules using the NumPy C API.

The following sections focus on the Numpy features supported in nopython mode, unless otherwise stated.

Scalar types

Numba supports the following Numpy scalar types:

• Integers: all integers of either signedness, and any width up to 64 bits
• Booleans
• Real numbers: single-precision (32-bit) and double-precision (64-bit) reals
• Complex numbers: single-precision (2x32-bit) and double-precision (2x64-bit) complex numbers
• Datetimes and timestamps: of any unit
• Character sequences (but no operations are available on them)
• Structured scalars: structured scalars made of any of the types above and arrays of the types above

The following scalar types and features are not supported:

• Arbitrary Python objects
• Half-precision and extended-precision real and complex numbers
• Nested structured scalars the fields of structured scalars may not contain other structured scalars

The operations supported on NumPy scalars are almost the same as on the equivalent built-in types such as int or float. You can use a type's constructor to convert from a different type or width. In addition you can use the view(np.<dtype>) method to bitcast all int and float types within the same width. However, you must define the scalar using a NumPy constructor within a jitted function. For example, the following will work:

>>> import numpy as np
>>> from numba import njit
>>> @njit
... def bitcast():
...     i = np.int64(-1)
...     print(i.view(np.uint64))
...
>>> bitcast()
18446744073709551615

Whereas the following will not work:

>>> import numpy as np
>>> from numba import njit
>>> @njit
... def bitcast(i):
...     print(i.view(np.uint64))
...
>>> bitcast(np.int64(-1))
---------------------------------------------------------------------------
TypingError                               Traceback (most recent call last)
    ...
TypingError: Failed in nopython mode pipeline (step: ensure IR is legal prior to lowering)
'view' can only be called on NumPy dtypes, try wrapping the variable with 'np.<dtype>()'

File "<ipython-input-3-fc40aaab84c4>", line 3:
def bitcast(i):
    print(i.view(np.uint64))

Structured scalars support attribute getting and setting, as well as member lookup using constant strings. Strings stored in a local or global tuple are considered constant strings and can be used for member lookup.

../../../numba/tests/doc_examples/test_rec_array.py

It is also possible to use local or global tuples together with literal_unroll:

../../../numba/tests/doc_examples/test_rec_array.py

Record subtyping

Warning

This is an experimental feature.

Numba allows width subtyping of structured scalars. For example, dtype([('a', 'f8'), ('b', 'i8')]) will be considered a subtype of dtype([('a', 'f8')], because the second is a strict subset of the first, i.e. field a is of the same type and is in the same position in both types. The subtyping relationship will matter in cases where compilation for a certain input is not allowed, but the input is a subtype of another, allowed type.

import numpy as np
from numba import njit, typeof
from numba.core import types
record1 = np.array([1], dtype=[('a', 'f8')])[0]
record2 = np.array([(2,3)], dtype=[('a', 'f8'), ('b', 'f8')])[0]

@njit(types.float64(typeof(record1)))
def foo(rec):
    return rec['a']

foo(record1)
foo(record2)

Without subtyping the last line would fail. With subtyping, no new compilation will be triggered, but the compiled function for record1 will be used for record2.

Numpy scalars reference.

Array types

Numpy arrays of any of the scalar types above are supported, regardless of the shape or layout.

Array access

Arrays support normal iteration. Full basic indexing and slicing is supported. A subset of advanced indexing is also supported: only one advanced index is allowed, and it has to be a one-dimensional array (it can be combined with an arbitrary number of basic indices as well).

Numpy indexing reference.

Structured array access

Numba presently supports accessing fields of individual elements in structured arrays by attribute as well as by getting and setting. This goes slightly beyond the NumPy API, which only allows accessing fields by getting and setting. For example:

from numba import njit
import numpy as np

record_type = np.dtype([("ival", np.int32), ("fval", np.float64)], align=True)

def f(rec):
    value = 2.5
    rec[0].ival = int(value)
    rec[0].fval = value
    return rec

arr = np.ones(1, dtype=record_type)

cfunc = njit(f)

# Works
print(cfunc(arr))

# Does not work
print(f(arr))

The above code results in the output:

[(2, 2.5)]
Traceback (most recent call last):
  File "repro.py", line 22, in <module>
    print(f(arr))
  File "repro.py", line 9, in f
    rec[0].ival = int(value)
AttributeError: 'numpy.void' object has no attribute 'ival'

The Numba-compiled version of the function executes, but the pure Python version raises an error because of the unsupported use of attribute access.

Note

This behavior will eventually be deprecated and removed.

Attributes

The following attributes of Numpy arrays are supported:

• ~numpy.ndarray.dtype
• ~numpy.ndarray.flags
• ~numpy.ndarray.flat
• ~numpy.ndarray.itemsize
• ~numpy.ndarray.ndim
• ~numpy.ndarray.shape
• ~numpy.ndarray.size
• ~numpy.ndarray.strides
• ~numpy.ndarray.T
• ~numpy.ndarray.real
• ~numpy.ndarray.imag

The flags object

The object returned by the ~numpy.ndarray.flags attribute supports the contiguous, c_contiguous and f_contiguous attributes.

The flat object

The object returned by the ~numpy.ndarray.flat attribute supports iteration and indexing, but be careful: indexing is very slow on non-C-contiguous arrays.

The real and imag attributes

Numpy supports these attributes regardless of the dtype but Numba chooses to limit their support to avoid potential user error. For numeric dtypes, Numba follows Numpy's behavior. The ~numpy.ndarray.real attribute returns a view of the real part of the complex array and it behaves as an identity function for other numeric dtypes. The ~numpy.ndarray.imag attribute returns a view of the imaginary part of the complex array and it returns a zero array with the same shape and dtype for other numeric dtypes. For non-numeric dtypes, including all structured/record dtypes, using these attributes will result in a compile-time (TypingError) error. This behavior differs from Numpy's but it is chosen to avoid the potential confusion with field names that overlap these attributes.

Calculation

The following methods of Numpy arrays are supported in their basic form (without any optional arguments):

• ~numpy.ndarray.all
• ~numpy.ndarray.any
• ~numpy.ndarray.argmax
• ~numpy.ndarray.argmin
• ~numpy.ndarray.conj
• ~numpy.ndarray.conjugate
• ~numpy.ndarray.cumprod
• ~numpy.ndarray.cumsum
• ~numpy.ndarray.max
• ~numpy.ndarray.mean
• ~numpy.ndarray.min
• ~numpy.ndarray.nonzero
• ~numpy.ndarray.prod
• ~numpy.ndarray.std
• ~numpy.ndarray.take
• ~numpy.ndarray.var

The corresponding top-level Numpy functions (such as numpy.prod) are similarly supported.

Other methods

The following methods of Numpy arrays are supported:

• ~numpy.ndarray.argsort (kind key word argument supported for values 'quicksort' and 'mergesort')
• ~numpy.ndarray.astype (only the 1-argument form)
• ~numpy.ndarray.copy (without arguments)
• ~numpy.ndarray.dot (only the 1-argument form)
• ~numpy.ndarray.flatten (no order argument; 'C' order only)
• ~numpy.ndarray.item (without arguments)
• ~numpy.ndarray.itemset (only the 1-argument form)
• ~numpy.ndarray.ptp (without arguments)
• ~numpy.ndarray.ravel (no order argument; 'C' order only)
• ~numpy.ndarray.repeat (no axis argument)
• ~numpy.ndarray.reshape (only the 1-argument form)
• ~numpy.ndarray.sort (without arguments)
• ~numpy.ndarray.sum (with or without the axis and/or dtype arguments.)
  • axis only supports integer values.
  • If the axis argument is a compile-time constant, all valid values are supported. An out-of-range value will result in a LoweringError at compile-time.
  • If the axis argument is not a compile-time constant, only values from 0 to 3 are supported. An out-of-range value will result in a runtime exception.
  • All numeric dtypes are supported in the dtype parameter. timedelta arrays can be used as input arrays but timedelta is not supported as dtype parameter.
  • When a dtype is given, it determines the type of the internal accumulator. When it is not, the selection is made automatically based on the input array's dtype, mostly following the same rules as NumPy. However, on 64-bit Windows, Numba uses a 64-bit accumulator for integer inputs (int64 for int32 inputs and uint64 for uint32 inputs), while NumPy would use a 32-bit accumulator in those cases.
• ~numpy.ndarray.transpose
• ~numpy.ndarray.view (only the 1-argument form)
• ~numpy.ndarray.__contains__

Warning

Sorting may be slightly slower than Numpy's implementation.

Functions

Linear algebra

Basic linear algebra is supported on 1-D and 2-D contiguous arrays of floating-point and complex numbers:

• numpy.dot
• numpy.kron ('C' and 'F' order only)
• numpy.outer
• numpy.trace (only the first argument).
• numpy.vdot
• On Python 3.5 and above, the matrix multiplication operator from 465 (i.e. a @ b where a and b are 1-D or 2-D arrays).
• numpy.linalg.cholesky
• numpy.linalg.cond (only non string values in p).
• numpy.linalg.det
• numpy.linalg.eig (only running with data that does not cause a domain change is supported e.g. real input -> real output, complex input -> complex output).
• numpy.linalg.eigh (only the first argument).
• numpy.linalg.eigvals (only running with data that does not cause a domain change is supported e.g. real input -> real output, complex input -> complex output).
• numpy.linalg.eigvalsh (only the first argument).
• numpy.linalg.inv
• numpy.linalg.lstsq
• numpy.linalg.matrix_power
• numpy.linalg.matrix_rank
• numpy.linalg.norm (only the 2 first arguments and only non string values in ord).
• numpy.linalg.pinv
• numpy.linalg.qr (only the first argument).
• numpy.linalg.slogdet
• numpy.linalg.solve
• numpy.linalg.svd (only the 2 first arguments).

Note

The implementation of these functions needs SciPy to be installed.

Reductions

The following reduction functions are supported:

• numpy.diff (only the 2 first arguments)
• numpy.median (only the first argument)
• numpy.nancumprod (only the first argument, requires NumPy >= 1.12))
• numpy.nancumsum (only the first argument, requires NumPy >= 1.12))
• numpy.nanmax (only the first argument)
• numpy.nanmean (only the first argument)
• numpy.nanmedian (only the first argument)
• numpy.nanmin (only the first argument)
• numpy.nanpercentile (only the 2 first arguments, requires NumPy >= 1.11, complex dtypes unsupported)
• numpy.nanquantile (only the 2 first arguments, requires NumPy >= 1.15, complex dtypes unsupported)
• numpy.nanprod (only the first argument)
• numpy.nanstd (only the first argument)
• numpy.nansum (only the first argument)
• numpy.nanvar (only the first argument)
• numpy.percentile (only the 2 first arguments, requires NumPy >= 1.10, complex dtypes unsupported)
• numpy.quantile (only the 2 first arguments, requires NumPy >= 1.15, complex dtypes unsupported)

Other functions

The following top-level functions are supported:

• numpy.append
• numpy.arange
• numpy.argsort (kind key word argument supported for values 'quicksort' and 'mergesort')
• numpy.argwhere
• numpy.array (only the 2 first arguments)
• numpy.array_equal
• numpy.array_split
• numpy.asarray (only the 2 first arguments)
• numpy.asarray_chkfinite (only the 2 first arguments)
• numpy.asfarray
• numpy.asfortranarray (only the first argument)
• numpy.atleast_1d
• numpy.atleast_2d
• numpy.atleast_3d
• numpy.bartlett
• numpy.bincount
• numpy.blackman
• numpy.column_stack
• numpy.concatenate
• numpy.convolve (only the 2 first arguments)
• numpy.copy (only the first argument)
• numpy.corrcoef (only the 3 first arguments, requires SciPy)
• numpy.correlate (only the 2 first arguments)
• numpy.count_nonzero (axis only supports scalar values)
• numpy.cov (only the 5 first arguments)
• numpy.cross (only the 2 first arguments; at least one of the input arrays should have shape[-1] == 3)
  • If shape[-1] == 2 for both inputs, please replace your numpy.cross call with numba.np.extensions.cross2d.
• numpy.delete (only the 2 first arguments)
• numpy.diag
• numpy.digitize
• numpy.dstack
• numpy.dtype (only the first argument)
• numpy.ediff1d
• numpy.empty (only the 2 first arguments)
• numpy.empty_like (only the 2 first arguments)
• numpy.expand_dims
• numpy.extract
• numpy.eye
• numpy.fill_diagonal
• numpy.flatten (no order argument; 'C' order only)
• numpy.flatnonzero
• numpy.flip (no axis argument)
• numpy.fliplr
• numpy.flipud
• numpy.frombuffer (only the 2 first arguments)
• numpy.full (only the 3 first arguments)
• numpy.full_like (only the 3 first arguments)
• numpy.hamming
• numpy.hanning
• numpy.histogram (only the 3 first arguments)
• numpy.hstack
• numpy.identity
• numpy.kaiser
• numpy.interp (only the 3 first arguments; requires NumPy >= 1.10)
• numpy.intersect1d (only first 2 arguments, ar1 and ar2)
• numpy.linspace (only the 3-argument form)
• numpy.ndenumerate
• numpy.ndindex
• numpy.nditer (only the first argument)
• numpy.ones (only the 2 first arguments)
• numpy.ones_like (only the 2 first arguments)
• numpy.partition (only the 2 first arguments)
• numpy.ptp (only the first argument)
• numpy.ravel (no order argument; 'C' order only)
• numpy.repeat (no axis argument)
• numpy.reshape (no order argument; 'C' order only)
• numpy.roll (only the 2 first arguments; second argument shift must be an integer)
• numpy.roots
• numpy.round_
• numpy.searchsorted (only the 3 first arguments)
• numpy.select (only using homogeneous lists or tuples for the first two arguments, condlist and choicelist). Additionally, these two arguments can only contain arrays (unlike Numpy that also accepts tuples).
• numpy.shape
• numpy.sinc
• numpy.sort (no optional arguments)
• numpy.split
• numpy.stack
• numpy.take (only the 2 first arguments)
• numpy.transpose
• numpy.trapz (only the 3 first arguments)
• numpy.tri (only the 3 first arguments; third argument k must be an integer)
• numpy.tril (second argument k must be an integer)
• numpy.tril_indices (all arguments must be integer)
• numpy.tril_indices_from (second argument k must be an integer)
• numpy.triu (second argument k must be an integer)
• numpy.triu_indices (all arguments must be integer)
• numpy.triu_indices_from (second argument k must be an integer)
• numpy.unique (only the first argument)
• numpy.vander
• numpy.vstack
• numpy.where
• numpy.zeros (only the 2 first arguments)
• numpy.zeros_like (only the 2 first arguments)

The following constructors are supported, both with a numeric input (to construct a scalar) or a sequence (to construct an array):

• numpy.bool_
• numpy.complex64
• numpy.complex128
• numpy.float32
• numpy.float64
• numpy.int8
• numpy.int16
• numpy.int32
• numpy.int64
• numpy.intc
• numpy.intp
• numpy.uint8
• numpy.uint16
• numpy.uint32
• numpy.uint64
• numpy.uintc
• numpy.uintp

The following machine parameter classes are supported, with all purely numerical attributes:

• numpy.iinfo
• numpy.finfo (machar attribute not supported)
• numpy.MachAr (with no arguments to the constructor)

Literal arrays

Neither Python nor Numba has actual array literals, but you can construct arbitrary arrays by calling numpy.array on a nested tuple:

a = numpy.array(((a, b, c), (d, e, f)))

(nested lists are not yet supported by Numba)

Modules

random

Numba supports top-level functions from the numpy.random module, but does not allow you to create individual RandomState instances. The same algorithms are used as for the standard random module <pysupported-random> (and therefore the same notes apply), but with an independent internal state: seeding or drawing numbers from one generator won't affect the other.

The following functions are supported.

Initialization

• numpy.random.seed: with an integer argument only

Simple random data

• numpy.random.rand
• numpy.random.randint (only the first two arguments)
• numpy.random.randn
• numpy.random.random
• numpy.random.random_sample
• numpy.random.ranf
• numpy.random.sample

Permutations

• numpy.random.choice: the optional p argument (probabilities array) is not supported
• numpy.random.permutation
• numpy.random.shuffle: the sequence argument must be a one-dimension Numpy array or buffer-providing object (such as a bytearray or array.array)

Distributions

Warning

The size argument is not supported in the following functions.

• numpy.random.beta
• numpy.random.binomial
• numpy.random.chisquare
• numpy.random.exponential
• numpy.random.f
• numpy.random.gamma
• numpy.random.geometric
• numpy.random.gumbel
• numpy.random.hypergeometric
• numpy.random.laplace
• numpy.random.logistic
• numpy.random.lognormal
• numpy.random.logseries
• numpy.random.multinomial
• numpy.random.negative_binomial
• numpy.random.normal
• numpy.random.pareto
• numpy.random.poisson
• numpy.random.power
• numpy.random.rayleigh
• numpy.random.standard_cauchy
• numpy.random.standard_exponential
• numpy.random.standard_gamma
• numpy.random.standard_normal
• numpy.random.standard_t
• numpy.random.triangular
• numpy.random.uniform
• numpy.random.vonmises
• numpy.random.wald
• numpy.random.weibull
• numpy.random.zipf

Note

Calling numpy.random.seed from non-Numba code (or from object mode code) will seed the Numpy random generator, not the Numba random generator.

Note

Since version 0.28.0, the generator is thread-safe and fork-safe. Each thread and each process will produce independent streams of random numbers.

stride_tricks

The following function from the numpy.lib.stride_tricks module is supported:

• ~numpy.lib.stride_tricks.as_strided (the strides argument is mandatory, the subok argument is not supported)

Standard ufuncs

One objective of Numba is having all the standard ufuncs in NumPy understood by Numba. When a supported ufunc is found when compiling a function, Numba maps the ufunc to equivalent native code. This allows the use of those ufuncs in Numba code that gets compiled in nopython mode.

Limitations

Right now, only a selection of the standard ufuncs work in nopython mode. Following is a list of the different standard ufuncs that Numba is aware of, sorted in the same way as in the NumPy documentation.

Math operations

UFUNC	MODE
name	object mode nopython mode
==============	============= ===============
add	Yes Yes
subtract	Yes Yes
multiply	Yes Yes
divide	Yes Yes
logaddexp	Yes Yes
logaddexp2	Yes Yes
true_divide	Yes Yes
floor_divide	Yes Yes
negative	Yes Yes
power	Yes Yes
remainder	Yes Yes
mod	Yes Yes
fmod	Yes Yes
divmod (*)	Yes Yes
abs	Yes Yes
absolute	Yes Yes
fabs	Yes Yes
rint	Yes Yes
sign	Yes Yes
conj	Yes Yes
exp	Yes Yes
exp2	Yes Yes
log	Yes Yes
log2	Yes Yes
log10	Yes Yes
expm1	Yes Yes
log1p	Yes Yes
sqrt	Yes Yes
square	Yes Yes
reciprocal	Yes Yes
conjugate	Yes Yes
gcd	Yes Yes
lcm	Yes Yes

(*) not supported on timedelta types

Trigonometric functions

UFUNC	MODE
name	object mode nopython mode
==============	============= ===============
sin	Yes Yes
cos	Yes Yes
tan	Yes Yes
arcsin	Yes Yes
arccos	Yes Yes
arctan	Yes Yes
arctan2	Yes Yes
hypot	Yes Yes
sinh	Yes Yes
cosh	Yes Yes
tanh	Yes Yes
arcsinh	Yes Yes
arccosh	Yes Yes
arctanh	Yes Yes
deg2rad	Yes Yes
rad2deg	Yes Yes
degrees	Yes Yes
radians	Yes Yes

Bit-twiddling functions

UFUNC	MODE
name	object mode nopython mode
==============	============= ===============
bitwise_and	Yes Yes
bitwise_or	Yes Yes
bitwise_xor	Yes Yes
bitwise_not	Yes Yes
invert	Yes Yes
left_shift	Yes Yes
right_shift	Yes Yes

Comparison functions

UFUNC	MODE
name	object mode nopython mode
==============	============= ===============
greater	Yes Yes
greater_equal	Yes Yes
less	Yes Yes
less_equal	Yes Yes
not_equal	Yes Yes
equal	Yes Yes
logical_and	Yes Yes
logical_or	Yes Yes
logical_xor	Yes Yes
logical_not	Yes Yes
maximum	Yes Yes
minimum	Yes Yes
fmax	Yes Yes
fmin	Yes Yes

Floating functions

UFUNC	MODE
name	object mode nopython mode
==============	============= ===============
isfinite	Yes Yes
isinf	Yes Yes
isnan	Yes Yes
signbit	Yes Yes
copysign	Yes Yes
nextafter	Yes Yes
modf	Yes No
ldexp	Yes (*) Yes
frexp	Yes No
floor	Yes Yes
ceil	Yes Yes
trunc	Yes Yes
spacing	Yes Yes

(*) not supported on windows 32 bit

Datetime functions

UFUNC	MODE
name	object mode nopython mode
==============	============= ===============
isnat	Yes Yes