NumPy ndarray.flags Attribute

NumPy ndarray.flags

The numpy.ndarray.flags attribute provides information about the memory layout of a NumPy array. It returns a dictionary-like object containing boolean flags indicating various memory properties of the array.

Syntax

</>

Copy

ndarray.flags

Parameters

This attribute does not take any parameters as it is a property of the ndarray object.

Return Value

Returns a numpy.flagsobj object containing boolean flags that provide information about the memory layout and other properties of the array.

Flag	Description
`C_CONTIGUOUS (C)`	Data is stored in a single C-style contiguous block.
`F_CONTIGUOUS (F)`	Data is stored in a single Fortran-style contiguous block.
`OWNDATA (O)`	Array owns the memory it uses or shares it with another array.
`WRITEABLE (W)`	Data can be written to.
`ALIGNED (A)`	Data and strides are aligned appropriately for the hardware.
`WRITEBACKIFCOPY (X)`	This array is a copy; writeback to the original array will occur upon destruction.
`UPDATEIFCOPY (U)`	This array is a copy; writeback to the original array will occur upon destruction (deprecated).
`FNC`	F-style contiguous or C-style contiguous.
`FORC`	F-style contiguous and C-style contiguous.

Examples

1. Checking Memory Layout Flags of a NumPy Array

</>

Copy

import numpy as np

# Creating a simple 2x2 array
arr = np.array([[1, 2], [3, 4]])

# Accessing the flags attribute
flags_info = arr.flags
print("Flags Information:\n", flags_info)

Output:

Flags Information:
   C_CONTIGUOUS : True
  F_CONTIGUOUS : False
  OWNDATA : True
  WRITEABLE : True
  ALIGNED : True
  WRITEBACKIFCOPY : False

The output indicates that the array is C-contiguous but not F-contiguous, owns its data, is writable, and aligned properly. The write-back flags are False, meaning no copy-on-write behavior is involved.

2. Checking Flags for a Transposed Array

</>

Copy

import numpy as np

# Creating a simple 2x3 array
arr = np.array([[1, 2, 3], [4, 5, 6]])

# Transposing the array
transposed_arr = arr.T

# Accessing the flags of the transposed array
transposed_flags = transposed_arr.flags
print("Flags of Transposed Array:\n", transposed_flags)

Output:

Flags of Transposed Array:
   C_CONTIGUOUS : False
  F_CONTIGUOUS : True
  OWNDATA : False
  WRITEABLE : True
  ALIGNED : True
  WRITEBACKIFCOPY : False

The transposed array is Fortran-contiguous but not C-contiguous, does not own its data (as it shares memory with the original array), and remains writable.

3. Creating a Non-Writeable Array

</>

Copy

import numpy as np

# Creating a simple array
arr = np.array([10, 20, 30])

# Setting the array to be non-writeable
arr.flags.writeable = False

# Trying to modify the array
try:
    arr[0] = 100
except ValueError as e:
    print("Error:", e)

# Checking the flags
print("Array Flags:\n", arr.flags)

Output:

Error: assignment destination is read-only
Array Flags:
   C_CONTIGUOUS : True
  F_CONTIGUOUS : True
  OWNDATA : True
  WRITEABLE : False
  ALIGNED : True
  WRITEBACKIFCOPY : False

The array has been set to non-writeable, indicated by the WRITEABLE flag being False. Attempting to modify the array raises a ValueError.

TutorialKart

NumPy ndarray.flags Attribute