在Python/Numpy中使用不同尺寸的阵列的最佳实践是什么？

发布于 2025-02-11 05:29:27 字数 380 浏览 1 评论 0原文

通常，您将如何使用numpy创建一个空数组并添加值。我正在撰写一个脚本，在该脚本中我迭代了一些数据，但是我不知道我需要添加的每个元素将要有多大，因此我无法预先分配内存。处理这种情况的最佳实践是什么？我可以将数据附加到对象类型数组中，但这很慢。有更好的方法来处理这一点吗？

示例：

# data is shape (100,256,256)
results_hist = np.zeros((0,0), dtype=object)
for i in data:
    results = doSomething(i) # varying size
    results_hist = np.append(results_hist,results)

这是最好的做法吗？

原文

Normally how you would use numpy is to create an empty array and add values in as they come. I am working on a script where I iterate over some data, but I don't know how large each element I need to add is going to be, so I'm not able to preallocate memory. What is the best practice for dealing with such a situation? I can append the data to a object type array, but that is slow. Is there a better way to handle this?

Example:

# data is shape (100,256,256)
results_hist = np.zeros((0,0), dtype=object)
for i in data:
    results = doSomething(i) # varying size
    results_hist = np.append(results_hist,results)

Is this the best practice?

分享到QQ

分享到微博