Compact integer arrays
Store nonnegative values in a compact, byte-sized representation.
NumPy · Type Alias
np.uint8
np.uint8 is NumPy’s unsigned 8-bit integer type. It represents nonnegative integer data and is commonly used as an array dtype for byte-sized values.
How it works
np.uint8 is NumPy’s unsigned integer scalar type with 8 bits and no sign. Use it as an array dtype or as a scalar constructor when values fit its representable range; NumPy identifies arrays using this dtype as uint8.
Syntax
numpy.uint8(value)Returns
A NumPy unsigned 8-bit integer scalar when called as np.uint8(value); when supplied as an array dtype, it specifies the array’s data type.
Key points
np.uint8stores unsigned integer values in 8 bits.- The canonical NumPy name is
numpy.uint8; its C-like alias isnumpy.ubyte, corresponding to an unsigned char. - When used as an array dtype,
np.uint8describes a NumPy data type with defined characteristics. - Convert an existing array to this dtype with
.astype(np.uint8). - Inspect an array’s
.dtypeattribute to confirm whether it usesuint8.
Common mistakes
- Overflow behavior differs between scalar and array operations: scalar overflow can warn, while equivalent array operations may not.
Example: Use uint8 for byte-sized unsigned values
import numpy as np
pixels = np.array([0, 128, 255], dtype=np.uint8)
print(pixels.dtype)
print(np.iinfo(np.uint8).min, np.iinfo(np.uint8).max)Creates an array with the unsigned 8-bit integer data type and displays its dtype and representable limits.
When to use it
Byte-oriented data
Represent image channels, masks, or other byte-oriented data.
Interoperability
Match interfaces that require an unsigned 8-bit integer.
Interview questions
QWhat does np.uint8 represent?
np.uint8 is an unsigned integer type stored using 8 bits. Its values have no sign, so the representable range is 0 through 255.
Follow-up: How would you verify the range programmatically? Use np.iinfo(np.uint8).min and .max rather than hard-coding limits.
QHow is np.uint8 commonly used?
Specify it as an array dtype, for example np.array([1, 2], dtype=np.uint8), or construct a NumPy scalar with np.uint8(1). The resulting array reports dtype('uint8').
Follow-up: How can an existing array be converted? Use its .astype(np.uint8) method, while ensuring the values are appropriate for the target type.
QWhat should engineers watch for during arithmetic?
Its 8-bit unsigned representation is useful when storage or interoperability requires byte-sized nonnegative integers. Arithmetic still follows NumPy dtype-promotion rules, so combining it with other types can change the result dtype.
Follow-up: What should be checked when mixing it with Python integers? Check the resulting dtype and possible overflow, because promotion and overflow behavior depend on the operands.
Related
numpy.ubytenumpy.dtypenumpy.iinfondarray.astypenumpy.uint16numpy.int8