Bing地图绘图工具模块(Bing Maps Drawing Tools Module)

 第二个问题是指针的问题，委托给C或其他语言不是解决方案，因为指针总是需要8或4个字节（取决于架构）。 另外，我假设使用64b python进行数十GB的工作，因此几乎每个变量类型都会大约2倍。 
 
 例如： 
 
node = [None, None, None, None]
node = [[[None, None, None, None], None, None, None], None, None, None]
 
 每个“None”都可以引用另一个节点（指针），因此不可能以这种方式进行优化。 如果您确定叶子的数量将少于~65K，那么您可以减少（~4次）并根据需要修改64b架构上的指针大小。 
 
 第一个问题的问题类似（指针总是占用必要的空间），解决方案是使用位数组或使用位操作，并尝试在一个变量中存储更多值。 但是如果有必要在某些结构中精确地存储例如2个字节的值然后例如在数组中，则在python中它将不具有存储器效率。 

Second problem is matter of pointers, delegate to C or another language is not solution because pointer will always take 8 or 4 bytes (depends on architecture). In addition, I assume to use 64b python for work with tens of GB so almost every variable type will be ~2 times greater.
 
For example:
 
node = [None, None, None, None]
node = [[[None, None, None, None], None, None, None], None, None, None]
 
Each "None" is possible reference to another node (pointer), so it is not possible to do optimization in this way. If you are sure that count of leaf will be e.g. less then ~65K, then you can decrease (~4 times) and modify size of pointer on 64b architecture for your needs.
 
Problem of first problem is similar (pointer will always take necessary space) and solution is use bit arrays or use bit operations and try to store more values in one variable. But if it is necessary to store exactly e.g 2 bytes values in some structure and then e.g. in array, it won't be memory-efficient in python.