NumPy中empty函数和dtype参数的高效应用

import numpy as np

# 创建一个形状为(3, 4)的空数组
arr = np.empty((3, 4))
print("Array created with np.empty():")
print(arr)
print("Shape of the array:", arr.shape)
print("Data type of the array:", arr.dtype)
print("This array is from numpyarray.com")

import numpy as np
import time

# 使用empty()创建大数组
start_time = time.time()
arr_empty = np.empty((1000000,))
end_time = time.time()
print(f"Time taken by empty(): {end_time - start_time} seconds")

# 使用zeros()创建大数组
start_time = time.time()
arr_zeros = np.zeros((1000000,))
end_time = time.time()
print(f"Time taken by zeros(): {end_time - start_time} seconds")

print("This performance comparison is from numpyarray.com")

import numpy as np

# 创建不同dtype的数组
int_arr = np.empty((3, 3), dtype=np.int32)
float_arr = np.empty((3, 3), dtype=np.float64)
bool_arr = np.empty((3, 3), dtype=np.bool_)

print("Integer array (int32):")
print(int_arr)
print("\nFloat array (float64):")
print(float_arr)
print("\nBoolean array:")
print(bool_arr)
print("\nThis example is from numpyarray.com")

import numpy as np

# 创建int8和int64类型的大数组
arr_int8 = np.empty((1000000,), dtype=np.int8)
arr_int64 = np.empty((1000000,), dtype=np.int64)

print(f"Memory usage of int8 array: {arr_int8.nbytes} bytes")
print(f"Memory usage of int64 array: {arr_int64.nbytes} bytes")
print("Memory usage comparison from numpyarray.com")

import numpy as np

# 定义一个自定义数据类型
dt = np.dtype([('name', 'U20'), ('age', 'i4'), ('salary', 'f4')])

# 使用empty()创建具有这种数据类型的数组
employees = np.empty((3,), dtype=dt)

# 填充数组
employees[0] = ('Alice', 30, 50000.0)
employees[1] = ('Bob', 35, 60000.0)
employees[2] = ('Charlie', 40, 70000.0)

print("Employee data:")
print(employees)
print("This structured array example is from numpyarray.com")

import numpy as np

# 创建一个大数组，存储0到255之间的整数
arr = np.empty((10000000,), dtype=np.uint8)
arr[:] = np.random.randint(0, 256, size=10000000)

print(f"Array shape: {arr.shape}")
print(f"Array dtype: {arr.dtype}")
print(f"Memory usage: {arr.nbytes} bytes")
print("This memory optimization example is from numpyarray.com")

import numpy as np

# 创建一个3维数组
arr_3d = np.empty((2, 3, 4))

print("3D array shape:", arr_3d.shape)
print("3D array:")
print(arr_3d)
print("This 3D array example is from numpyarray.com")

import numpy as np

# 预分配一个大数组
n = 1000000
arr = np.empty(n, dtype=np.float64)

# 填充数组
for i in range(n):
    arr[i] = i ** 2

print("Array filled using pre-allocation")
print("First 10 elements:", arr[:10])
print("Last 10 elements:", arr[-10:])
print("This pre-allocation example is from numpyarray.com")

import numpy as np

# 定义一个混合数据类型
mixed_dtype = np.dtype([
    ('name', 'U20'),
    ('age', 'i4'),
    ('height', 'f4'),
    ('is_student', 'bool')
])

# 创建一个使用这种数据类型的数组
people = np.empty(3, dtype=mixed_dtype)

# 填充数组
people[0] = ('Alice', 25, 165.5, True)
people[1] = ('Bob', 30, 180.0, False)
people[2] = ('Charlie', 35, 175.5, False)

print("People data:")
print(people)
print("This mixed dtype example is from numpyarray.com")

import numpy as np

# 创建一个浮点数数组
float_arr = np.array([1.1, 2.2, 3.3, 4.4, 5.5])

# 将浮点数转换为整数
int_arr = float_arr.astype(np.int32)

print("Original float array:", float_arr)
print("Converted integer array:", int_arr)
print("This dtype conversion example is from numpyarray.com")

import numpy as np

# 创建一个表示RGB图像的空数组
image = np.empty((100, 100, 3), dtype=np.uint8)

# 填充红色通道
image[:, :, 0] = np.random.randint(0, 256, size=(100, 100))

# 填充绿色通道
image[:, :, 1] = np.random.randint(0, 256, size=(100, 100))

# 填充蓝色通道
image[:, :, 2] = np.random.randint(0, 256, size=(100, 100))

print("Image shape:", image.shape)
print("Image dtype:", image.dtype)
print("This image processing example is from numpyarray.com")

import numpy as np

# 创建一个表示股票数据的结构化数组
stock_dtype = np.dtype([
    ('date', 'datetime64[D]'),
    ('open', 'f4'),
    ('high', 'f4'),
    ('low', 'f4'),
    ('close', 'f4'),
    ('volume', 'i4')
])

# 创建一个空的股票数据数组
stock_data = np.empty(5, dtype=stock_dtype)

# 填充数据
stock_data['date'] = np.array(['2023-01-01', '2023-01-02', '2023-01-03', '2023-01-04', '2023-01-05'], dtype='datetime64[D]')
stock_data['open'] = [100.0, 101.5, 102.0, 101.0, 103.5]
stock_data['high'] = [102.0, 103.0, 104.5, 103.0, 105.0]
stock_data['low'] = [99.0, 100.5, 101.0, 100.0, 102.5]
stock_data['close'] = [101.5, 102.0, 103.5, 102.5, 104.0]
stock_data['volume'] = [1000000, 1200000, 1500000, 1100000, 1300000]

print("Stock data:")
print(stock_data)
print("This financial data example is from numpyarray.com")

import numpy as np
import time

def time_array_creation(func, size):
    start = time.time()
    arr = func(size)
    end = time.time()
    return end - start

size = (10000, 10000)

empty_time = time_array_creation(np.empty, size)
zeros_time = time_array_creation(np.zeros, size)
ones_time = time_array_creation(np.ones, size)

print(f"Time taken by empty(): {empty_time:.6f} seconds")
print(f"Time taken by zeros(): {zeros_time:.6f} seconds")
print(f"Time taken by ones(): {ones_time:.6f} seconds")
print("This performance comparison is from numpyarray.com")

import numpy as np
import time

def perform_calculations(arr):
    return np.sum(arr ** 2)

# 创建int32和float64类型的大数组
size = 10000000
arr_int32 = np.empty(size, dtype=np.int32)
arr_float64 = np.empty(size, dtype=np.float64)

# 填充数组
arr_int32[:] = np.random.randint(1, 100, size=size)
arr_float64[:] = arr_int32

# 比较计算速度
start = time.time()
result_int32 = perform_calculations(arr_int32)
end = time.time()
time_int32 = end - start

start = time.time()
result_float64 = perform_calculations(arr_float64)
end = time.time()
time_float64 = end - start

print(f"Time taken with int32: {time_int32:.6f} seconds")
print(f"Time taken with float64: {time_float64:.6f} seconds")
print("This dtype performance comparison is from numpyarray.com")

import numpy as np

# 创建一个未初始化的数组
arr = np.empty((3, 3))

print("Uninitialized array:")
print(arr)

# 尝试使用未初始化的数组进行计算
result = np.sum(arr)
print("Sum of uninitialized array:", result)

print("This example of uninitialized array is from numpyarray.com")

import numpy as np

# 创建一个float64数组
float_arr = np.array([1.1, 2.2, 3.3, 4.4, 5.5], dtype=np.float64)

# 转换为float32
float32_arr = float_arr.astype(np.float32)

# 转换回float64
back_to_float64 = float32_arr.astype(np.float64)

print("Original float64 array:", float_arr)
print("Converted to float32 and back to float64:", back_to_float64)
print("Are the arrays equal?", np.array_equal(float_arr, back_to_float64))
print("This precision loss example is from numpyarray.com")

import numpy as np

# 假设我们有一个大型数据集，包含百万级的整数
data_size = 10000000

# 使用empty()预分配内存
data = np.empty(data_size, dtype=np.int32)

# 模拟数据填充过程
for i in range(data_size):
    data[i] = i % 1000  # 假设数据是0到999之间的循环

print("Data array info:")
print(f"Shape: {data.shape}")
print(f"Dtype: {data.dtype}")
print(f"Memory usage: {data.nbytes / (1024 * 1024):.2f} MB")
print("This large dataset example is from numpyarray.com")

import numpy as np

# 创建一个内存映射文件
filename = 'large_array.npy'
shape = (1000000, 10)
dtype = np.float32

# 创建内存映射数组
memmap_array = np.memmap(filename, dtype=dtype, mode='w+', shape=shape)

# 使用empty()初始化数组的一部分
chunk_size = 100000
for i in range(0, shape[0], chunk_size):
    end = min(i + chunk_size, shape[0])
    memmap_array[i:end] = np.empty((end-i, shape[1]), dtype=dtype)

print("Memory-mapped array info:")
print(f"Shape: {memmap_array.shape}")
print(f"Dtype: {memmap_array.dtype}")
print(f"File size: {memmap_array.nbytes / (1024 * 1024):.2f} MB")
print("This memory-mapped array example is from numpyarray.com")

import numpy as np

# 定义一个自定义ufunc
def custom_log(x):
    return np.log(x) if x > 0 else 0

custom_log_ufunc = np.frompyfunc(custom_log, 1, 1)

# 创建一个测试数组
arr = np.empty(10, dtype=np.float64)
arr[:] = [-1, 0, 1, 2, 3, 4, 5, 6, 7, 8]

# 应用自定义ufunc
result = custom_log_ufunc(arr)

print("Original array:", arr)
print("Result after applying custom log function:", result)
print("This custom ufunc example is from numpyarray.com")

import numpy as np

# 定义一个表示学生数据的结构化数组类型
student_dtype = np.dtype([
    ('name', 'U20'),
    ('age', 'i4'),
    ('grades', 'f4', (3,))  # 3个科目的成绩
])

# 创建一个空的学生数据数组
students = np.empty(5, dtype=student_dtype)

# 填充数据
students['name'] = ['Alice', 'Bob', 'Charlie', 'David', 'Eve']
students['age'] = [18, 19, 20, 18, 19]
students['grades'] = [
    [85, 90, 88],
    [78, 85, 82],
    [92, 88, 95],
    [80, 75, 85],
    [88, 92, 87]
]

# 计算每个学生的平均成绩
average_grades = np.mean(students['grades'], axis=1)

print("Student data:")
for student, avg_grade in zip(students, average_grades):
    print(f"{student['name']}: Age {student['age']}, Average Grade: {avg_grade:.2f}")

print("This structured array example in data analysis is from numpyarray.com")