Scalars — NumPy v2.6.dev0 Manual
Python defines only one type of a particular data class (there is only one integer type, one floating-point type, etc.). This can be convenient in applications that don’t need to be concerned with all the ways data can be represented in a computer. For scientific computing, however, more control is often needed.
In NumPy, there are 24 new fundamental Python types to describe different types of scalars. These type descriptors are mostly based on the types available in the C language that CPython is written in, with several additional types compatible with Python’s types.
Array scalars have the same attributes and methods as
.
This allows one to treat items of an array partly on the same footing as arrays, smoothing out rough edges that result when mixing scalar and array operations.
Array scalars live in a hierarchy (see the Figure below) of data types. They can be detected using the hierarchy: For example, isinstance(val,np.generic) will return
if val is an array scalar object. Alternatively, what kind of array scalar is present can be determined using other members of the data type hierarchy. Thus, for example isinstance(val,np.complexfloating) will return
if val is a complex valued type, while isinstance(val,np.flexible) will return true if val is one of the flexible itemsize array types (
,
,
).
Figure: Hierarchy of type objects representing the array data types. Not shown are the two integer types
and
which are used for indexing (the same as the default integer since NumPy 2).
Built-in scalar types
The built-in scalar types are shown below. The C-like names are associated with character codes, which are shown in their descriptions. Use of the character codes, however, is discouraged.
Some of the scalar types are essentially equivalent to fundamental Python types and therefore inherit from them as well as from the generic array scalar type:
Array scalar type
Related Python type
Inherits?
no
yes
yes
yes
yes
no
no
no
The
data type is very similar to the Python
but does not inherit from it because Python’s
does not allow itself to be inherited from, and on the C-level the size of the actual bool data is not the same as a Python Boolean scalar.
Warning
The
type does not inherit from the built-in
, because type
is not a fixed-width integer type.
Tip
The default data type in NumPy is
.
classnumpy.generic
Base class for numpy scalar types.
Class from which most (all?) numpy scalar types are derived. For consistency, exposes the same API as
, despite many consequent attributes being either “get-only,” or completely irrelevant. This is the class from which it is strongly suggested users should derive custom scalar types.
classnumpy.number
Abstract base class of all numeric scalar types.
Integer types
classnumpy.integer
Abstract base class of all integer scalar types.
Note
The numpy integer types mirror the behavior of C integers, and can therefore be subject to
.
Signed integer types
classnumpy.signedinteger
Abstract base class of all signed integer scalar types.
classnumpy.byte
int8(value=0, /) –
Signed integer type, compatible with C char.
Character code:'b'
Canonical name:
Alias on this platform (Linux x86_64):
: 8-bit signed integer (-128 to 127).
classnumpy.short
int16(value=0, /) –
Signed integer type, compatible with C short.
Character code:'h'
Canonical name:
Alias on this platform (Linux x86_64):
: 16-bit signed integer (-32_768 to 32_767).
classnumpy.intc
int32(value=0, /) –
Signed integer type, compatible with C int.
Character code:'i'
Canonical name:
Alias on this platform (Linux x86_64):
: 32-bit signed integer (-2_147_483_648 to 2_147_483_647).
classnumpy.int_
int64(value=0, /) –
Signed integer type, compatible with C long.
Character code:'l'
Canonical name:
Alias on this platform (Linux x86_64):
: 64-bit signed integer (-9_223_372_036_854_775_808 to 9_223_372_036_854_775_807).
Alias on this platform (Linux x86_64):
: Signed integer large enough to fit pointer, compatible with C intptr_t.
numpy.long
alias of
classnumpy.longlong(value=0, /)
Signed integer type, compatible with C longlong.
Character code:'q'
Unsigned integer types
classnumpy.unsignedinteger
Abstract base class of all unsigned integer scalar types.
classnumpy.ubyte
uint8(value=0, /) –
Unsigned integer type, compatible with C unsignedchar.
Character code:'B'
Canonical name:
Alias on this platform (Linux x86_64):
: 8-bit unsigned integer (0 to 255).
classnumpy.ushort
uint16(value=0, /) –
Unsigned integer type, compatible with C unsignedshort.
Character code:'H'
Canonical name:
Alias on this platform (Linux x86_64):
: 16-bit unsigned integer (0 to 65_535).
classnumpy.uintc
uint32(value=0, /) –
Unsigned integer type, compatible with C unsignedint.
Character code:'I'
Canonical name:
Alias on this platform (Linux x86_64):
: 32-bit unsigned integer (0 to 4_294_967_295).
classnumpy.uint
uint64(value=0, /) –
Unsigned signed integer type, 64bit on 64bit systems and 32bit on 32bit systems.
Character code:'L'
Canonical name:
Alias on this platform (Linux x86_64):
: 64-bit unsigned integer (0 to 18_446_744_073_709_551_615).
Alias on this platform (Linux x86_64):
: Unsigned integer large enough to fit pointer, compatible with C uintptr_t.
numpy.ulong
alias of
classnumpy.ulonglong(value=0, /)
Unsigned integer type, compatible with C unsignedlonglong.
Character code:'Q'
Inexact types
classnumpy.inexact
Abstract base class of all numeric scalar types with a (potentially) inexact representation of the values in its range, such as floating-point numbers.
Note
Inexact scalars are printed using the fewest decimal digits needed to distinguish their value from other values of the same datatype, by judicious rounding. See the unique parameter of
and
.
This means that variables with equal binary values but whose datatypes are of different precisions may display differently:
>>> importnumpyasnp>>> f16=np.float16("0.1")>>> f32=np.float32(f16)>>> f64=np.float64(f32)>>> f16==f32==f64True>>> f16,f32,f64(0.1, 0.099975586, 0.0999755859375)Note that none of these floats hold the exact value \(\frac{1}{10}\); f16 prints as 0.1 because it is as close to that value as possible, whereas the other types do not as they have more precision and therefore have closer values.
Conversely, floating-point scalars of different precisions which approximate the same decimal value may compare unequal despite printing identically:
>>> f16=np.float16("0.1")>>> f32=np.float32("0.1")>>> f64=np.float64("0.1")>>> f16==f32==f64False>>> f16,f32,f64(0.1, 0.1, 0.1)Floating-point types
classnumpy.floating
Abstract base class of all floating-point scalar types.
classnumpy.half
float16(value=0, /) –
Half-precision floating-point number type.
Character code:'e'
Canonical name:
Alias on this platform (Linux x86_64):
: 16-bit-precision floating-point number type: sign bit, 5 bits exponent, 10 bits mantissa.
classnumpy.single
float32(value=0, /) –
Single-precision floating-point number type, compatible with C float.
Character code:'f'
Canonical name:
Alias on this platform (Linux x86_64):
: 32-bit-precision floating-point number type: sign bit, 8 bits exponent, 23 bits mantissa.
classnumpy.double(x=0, /)
float64(value=0, /) –
Double-precision floating-point number type, compatible with Python
and C double.
Character code:'d'
Canonical name:
Alias on this platform (Linux x86_64):
: 64-bit precision floating-point number type: sign bit, 11 bits exponent, 52 bits mantissa.
classnumpy.longdouble(value=0, /)
Extended-precision floating-point number type, compatible with C longdouble but not necessarily with IEEE 754 quadruple-precision.
Character code:'g'
Alias on this platform (Linux x86_64):
: 128-bit extended-precision floating-point number type.
Complex floating-point types
classnumpy.complexfloating
Abstract base class of all complex number scalar types that are made up of floating-point numbers.
classnumpy.csingle
complex64(real=0, imag=0, /) –
Complex number type composed of two single-precision floating-point numbers.
Character code:'F'
Canonical name:
Alias on this platform (Linux x86_64):
: Complex number type composed of 2 32-bit-precision floating-point numbers.
classnumpy.cdouble(real=0, imag=0)
complex128(real=0, imag=0, /) –
Complex number type composed of two double-precision floating-point numbers, compatible with Python
.
Character code:'D'
Canonical name:
Alias on this platform (Linux x86_64):
: Complex number type composed of 2 64-bit-precision floating-point numbers.
classnumpy.clongdouble(real=0, imag=0, /)
Complex number type composed of two extended-precision floating-point numbers.
Character code:'G'
Alias on this platform (Linux x86_64):
: Complex number type composed of 2 128-bit extended-precision floating-point numbers.
Other types
numpy.bool_
alias of
classnumpy.bool(value=False, /)
Boolean type (True or False), stored as a byte.
Warning
The
type is not a subclass of the
type (the
is not even a number type). This is different than Python’s default implementation of
as a sub-class of
.
Character code:'?'
classnumpy.datetime64(value=None, *args)
If created from a 64-bit integer, it represents an offset from 1970-01-01T00:00:00. If created from string, the string can be in ISO 8601 date or datetime format.
When parsing a string to create a datetime object, if the string contains a trailing timezone (A ‘Z’ or a timezone offset), the timezone will be dropped and a User Warning is given.
Datetime64 objects should be considered to be UTC and therefore have an offset of +0000.
>>> np.datetime64(10,'Y')np.datetime64('1980')>>> np.datetime64('1980','Y')np.datetime64('1980')>>> np.datetime64(10,'D')np.datetime64('1970-01-11')See
for more information.
Character code:'M'
classnumpy.timedelta64(value=0, *args)
A timedelta stored as a 64-bit integer.
See
for more information.
Character code:'m'
classnumpy.object_(value=None, /)
Any Python object.
Character code:'O'
Note
The data actually stored in object arrays (i.e., arrays having dtype
) are references to Python objects, not the objects themselves. Hence, object arrays behave more like usual Python
, in the sense that their contents need not be of the same Python type.
The object type is also special because an array containing
items does not return an
object on item access, but instead returns the actual object that the array item refers to.
The following data types are flexible: they have no predefined size and the data they describe can be of different length in different arrays. (In the character codes # is an integer denoting how many elements the data type consists of.)
classnumpy.flexible
Abstract base class of all scalar types without predefined length. The actual size of these types depends on the specific
instantiation.
classnumpy.character
Abstract base class of all character string scalar types.
classnumpy.bytes_(value='', *args, **kwargs)
A byte string.
When used in arrays, this type strips trailing null bytes.
Character code:'S'
classnumpy.str_(value='', *args, **kwargs)
A unicode string.
This type strips trailing null codepoints.
>>> s=np.str_("abc\x00")>>> s'abc'Unlike the builtin
, this supports the
, exposing its contents as UCS4:
>>> m=memoryview(np.str_("abc"))>>> m.format'3w'>>> m.tobytes()b'a\x00\x00\x00b\x00\x00\x00c\x00\x00\x00'Character code:'U'
classnumpy.void(length_or_data, /, dtype=None)
Create a new structured or unstructured void scalar.
Parameters:length_or_dataint, array-like, bytes-like, objectOne of multiple meanings (see notes). The length or bytes data of an unstructured void. Or alternatively, the data to be stored in the new scalar when
is provided. This can be an array-like, in which case an array may be returned.
dtypedtype, optionalIf provided the dtype of the new scalar. This dtype must be “void” dtype (i.e. a structured or unstructured void, see also
).
Added in version 1.24.
Notes
For historical reasons and because void scalars can represent both arbitrary byte data and structured dtypes, the void constructor has three calling conventions:
np.void(5) creates a dtype="V5" scalar filled with five\0 bytes. The 5 can be a Python or NumPy integer.
np.void(b"bytes-like") creates a void scalar from the byte string.The dtype itemsize will match the byte string length, here "V10".
When a dtype= is passed the call is roughly the same as anarray creation. However, a void scalar rather than array is returned.
Please see the examples which show all three different conventions.
Examples
>>> np.void(5)np.void(b'\x00\x00\x00\x00\x00')>>> np.void(b'abcd')np.void(b'\x61\x62\x63\x64')>>> np.void((3.2,b'eggs'),dtype="d,S5")np.void((3.2, b'eggs'), dtype=[('f0', '<f8'), ('f1', 'S5')])>>> np.void(3,dtype=[('x',np.int8),('y',np.int8)])np.void((3, 3), dtype=[('x', 'i1'), ('y', 'i1')])Character code:'V'
Warning
Numeric Compatibility: If you used old typecode characters in your Numeric code (which was never recommended), you will need to change some of them to the new characters. In particular, the needed changes are c->S1, b->B, 1->b, s->h, w->H, and u->I. These changes make the type character convention more consistent with other Python modules such as the
module.
Sized aliases
Along with their (mostly) C-derived names, the integer, float, and complex data-types are also available using a bit-width convention so that an array of the right size can always be ensured. Two aliases (
and
) pointing to the integer type that is sufficiently large to hold a C pointer are also provided.
numpy.int8
numpy.int16
numpy.int32
numpy.int64
Aliases for the signed integer types (one of
,
,
,
,
and
) with the specified number of bits.
Compatible with the C99 int8_t, int16_t, int32_t, and int64_t, respectively.
numpy.uint8
numpy.uint16
numpy.uint32
numpy.uint64
Alias for the unsigned integer types (one of
,
,
,
,
and
) with the specified number of bits.
Compatible with the C99 uint8_t, uint16_t, uint32_t, and uint64_t, respectively.
numpy.intp
Alias for the signed integer type (one of
,
,
,
,
and
) that is used as a default integer and for indexing.
Compatible with the C Py_ssize_t.
Character code:'n'
Changed in version 2.0: Before NumPy 2, this had the same size as a pointer. In practice this is almost always identical, but the character code 'p' maps to the C intptr_t. The character code 'n' was added in NumPy 2.0.
numpy.uintp
Alias for the unsigned integer type that is the same size as intp.
Compatible with the C size_t.
Character code:'N'
Changed in version 2.0: Before NumPy 2, this had the same size as a pointer. In practice this is almost always identical, but the character code 'P' maps to the C uintptr_t. The character code 'N' was added in NumPy 2.0.
numpy.float16
alias of
numpy.float32
alias of
numpy.float64
alias of
numpy.float96
numpy.float128
Alias for
, named after its size in bits. The existence of these aliases depends on the platform.
numpy.complex64
alias of
numpy.complex128
alias of
numpy.complex192
numpy.complex256
Alias for
, named after its size in bits. The existence of these aliases depends on the platform.
Attributes
The array scalar objects have an
of
(-1,000,000.0). They also do not (yet) have a
attribute. Otherwise, they share the same attributes as arrays:
The integer value of flags.
Tuple of array dimensions.
Tuple of bytes steps in each dimension.
The number of array dimensions.
Pointer to start of data.
The number of elements in the gentype.
The length of one element in bytes.
Scalar attribute identical to
.
Get array data-descriptor.
The real part of the scalar.
The imaginary part of the scalar.
A 1-D view of the scalar.
Scalar attribute identical to
.
Array protocol: Python side
Array protocol: struct
Array priority.
__array_wrap__ implementation for scalar types
Indexing
Array scalars can be indexed like 0-dimensional arrays: if x is an array scalar,
x[()] returns a copy of array scalar
x[...] returns a 0-dimensional
x['field-name'] returns the array scalar in the field field-name. (x can have fields, for example, when it corresponds to a structured data type.)
Methods
Array scalars have exactly the same methods as arrays. The default behavior of these methods is to internally convert the scalar to an equivalent 0-dimensional array and to call the corresponding array method. In addition, math operations on array scalars are defined so that the same hardware flags are set and used to interpret the results as for
, so that the error state used for ufuncs also carries over to the math on array scalars.
The exceptions to the above rules are given below:
sc.__array__(dtype) return 0-dim array from scalar with specified dtype
__array_wrap__ implementation for scalar types
(/[, axis])
Scalar method identical to
.
(/[, inplace])
Scalar method identical to
.
Helper for pickle.
(*[, write, align, uic])
Scalar method identical to
.
Utility method for typing:
Defining new types
There are two ways to effectively define a new array scalar type (apart from composing structured types
from the built-in scalar types): One way is to simply subclass the
and overwrite the methods of interest. This will work to a degree, but internally certain behaviors are fixed by the data type of the array. To fully customize the data type of an array you need to define a new data-type, and register it with NumPy. Such new types can only be defined in C, using the
.