In this appendix , I will go deep into NumPy Array calculation of library . This will include ndarray More internal details , And more advanced array operations and algorithms .

This chapter includes a few messy chapters , You don't need to study it carefully .

A.1 ndarray The internal mechanism of the object

NumPy Of ndarray It provides a way to integrate homogeneous data blocks （ It can be continuous or across ） How to interpret as a multidimensional array object . As you have seen before , data type （dtype） Determines how the data is interpreted , Floating point numbers, for example 、 Integers 、 Boolean value, etc .

ndarray Part of the reason why this is so powerful is that all array objects are a spanning view of data blocks （strided view）. You might want to know about array views arr[::2,::-1] Why not replicate any data . In short ,ndarray Not just a piece of memory and a dtype, It also has span information , This allows the array to move in various steps （step size） Move in memory . To be more precise ,ndarray The interior consists of ：

• A point to data （ A piece of data in a memory or memory mapped file ） The pointer to .
• Data type or dtype, A cell that describes a fixed size value in an array .
• An array shape （shape） tuples .
• A span tuple （stride）, The integer refers to the need to move forward to the next element of the current dimension “ Across ” Bytes of .

chart A-1 It simply explains ndarray The internal structure of .

for example , One 10×5 Array of , Its shape is (10,5)：