如何获取numpy dtype的字节序？以及判断numpy数组的实际字节序

Hey there! Let's tackle your two NumPy byteorder questions clearly and practically:

Getting the byteorder of a NumPy dtype is straightforward—every dtype object has a byteorder attribute you can access directly. Here are quick examples for different cases:

• For a little-endian dtype:

  import numpy as np
  little_endian_arr = np.array([1, 2, 3], dtype='<i4')
  print(little_endian_arr.dtype.byteorder)  # Output: '<'
  

• For a big-endian dtype:

  big_endian_arr = np.array([1, 2, 3], dtype='>i4')
  print(big_endian_arr.dtype.byteorder)  # Output: '>'
  

• For dtypes using your system's native byteorder, you'll get '=' (we'll handle this in the next question).
• For single-byte dtypes (like uint8 or int8), byteorder doesn't matter, so you'll see '|'.

The '=' symbol just means the dtype uses your system's native byteorder. To map this to a concrete </> symbol or a "大端"/"小端" conclusion, you can leverage Python's built-in sys module to check your system's native byteorder.

Option 1: Get the actual </> symbol #

import sys
import numpy as np

x = np.arange(3)
dtype_byteorder = x.dtype.byteorder

# Resolve '=' to system's actual byteorder symbol
actual_byteorder = '<' if sys.byteorder == 'little' else '>' if dtype_byteorder == '=' else dtype_byteorder

print(actual_byteorder)  # Outputs '<' or '>' based on your system

Option 2: Get a human-readable conclusion ("大端"/"小端"/"无关") #

import sys
import numpy as np

x = np.arange(3)
dtype_byteorder = x.dtype.byteorder

if dtype_byteorder == '|':
    print("无关")
elif dtype_byteorder == '=':
    print("小端" if sys.byteorder == 'little' else "大端")
else:
    print("小端" if dtype_byteorder == '<' else "大端")

A quick note: The '|' case applies to single-byte data types—since there's only one byte, byteorder doesn't affect how the data is stored, hence the "无关" conclusion.

内容的提问来源于stack exchange，提问作者Mad Physicist