一、数组与维度
从以上可知数组最前面有多少个[,就表示有多少维。
二、axis维度轴与数组与维度
备注:axis=0表示最高维度n轴,axis=1表示n-1维轴,以此类推
维度在数组中的理解,见下文:
>>> import numpy as np
>>> t = np.array(
... [
... [
... [
... [1,2,3],
... [4,5,6]
... ],
... [
... [7,8,9],
... [10,11,12]
... ],
... [
... [13,14,15],
... [16,17,18]
... ]
... ],
... [
... [
... [19,20,21],
... [22,23,24]
... ],
... [
... [25,26,27],
... [28,29,30]
... ],
... [
... [31,32,33],
... [34,35,36]
... ]
... ]
... ])
>>> print(t.shape)
(2, 3, 2, 3)
>>> print(t[0,:,:,:])#第0轴或者说第4维(也就是第一个[])下的第一项所有的数据
[[[ 1 2 3]
[ 4 5 6]]
[[ 7 8 9]
[10 11 12]]
[[13 14 15]
[16 17 18]]]
>>> print(t[:,0,:,:])#第axis=1轴或者第2个[]下说包含的第一项所有数据
[[[ 1 2 3]
[ 4 5 6]]
[[19 20 21]
[22 23 24]]]
>>> print(t[:,:,0,:])
[[[ 1 2 3]
[ 7 8 9]
[13 14 15]]
[[19 20 21]
[25 26 27]
[31 32 33]]]
>>> print(t[:,:,:,0])
[[[ 1 4]
[ 7 10]
[13 16]]
[[19 22]
[25 28]
[31 34]]]
三、np.concatenate(),np.expand_dims(),np.newaxis()
3.1、np.concatenate(对数列或矩阵进行合并)
一维情况:
>>> import numpy as np
>>> a=[1,2,3]
>>> b=[4,5,6]
>>> np.concatenate((a,b),axis=0) # 默认情况下,axis=0可以不写
[1, 2, 3, 4, 5, 6]
备注:因为上述a和b都是一维的,所以当指定axis=1时,程序就会报错。
二维情况:
>>> y1 = np.array([ [[1,0],[1,0]] , [[0,0],[0,0]] ])
>>> y2 = np.array([ [[0,0],[0,0]] , [[0,1],[0,1]] ])
>>> print(y1.shape)
>>> y3 = np.concatenate((y1,y2),axis=0)
# 可知y1,y2都是3维数组,axis=1则表示y1,y2的拼接方式按照3-0=3维的结构来拼接
>>> print(y3)
>>> print(y3.shape)
(2, 2, 2)
[[[1 0]
[1 0]]
[[0 0]
[0 0]]
[[0 0]
[0 0]]
[[0 1]
[0 1]]]
(4, 2, 2)
>>> y1 = np.array([ [[1,0],[1,0]] , [[0,0],[0,0]] ])
>>> y2 = np.array([ [[0,0],[0,0]] , [[0,1],[0,1]] ])
>>> print(y1.shape)
>>> y3 = np.concatenate((y1,y2),axis=1)
# 可知y1,y2都是3维数组,axis=1则表示y1,y2的拼接方式按照3-1=2维的结构来拼接
>>> print(y3)
>>> print(y3.shape)
(2, 2, 2)
[[[1 0]
[1 0]
[0 0]
[0 0]]
[[0 0]
[0 0]
[0 1]
[0 1]]]
(2, 4, 2)
>>> y1 = np.array([ [[1,0],[1,0]] , [[0,0],[0,0]] ])
>>> y2 = np.array([ [[0,0],[0,0]] , [[0,1],[0,1]] ])
>>> print(y1.shape)
>>> y3 = np.concatenate((y1,y2),axis=2)
# 可知y1,y2都是3维数组,axis=2则表示y1,y2的拼接方式按照3-2=1维的结构来拼接
>>> print(y3)
>>> print(y3.shape)
(2, 2, 2)
[[[1 0 0 0]
[1 0 0 0]]
[[0 0 0 1]
[0 0 0 1]]]
(2, 2, 4)
备注:所谓的按A维拼接,意思是两组数据对应的A维度数据相加,其他维度不改变(很关键,理解这句话和上面内容)
3.2、np.expand_dims(扩展数组维度)
一维情况:
>>> x = np.array([1,2])
>>> x.shape
(2,)
>>> y = np.expand_dims(x, axis=0) #扩充最高维,等价于 x[np.newaxis,:]或x[np.newaxis]
>>> y
array([[1, 2]])
>>> y.shape
(1, 2) #看np.newaxis位置(在:之前)可知插入在2之前
>>> y = np.expand_dims(x, axis=1) #等价于x[:,newaxis]
>>> y
array([[1],
[2]])
>>> y.shape
(2, 1) #看np.newaxis位置(在:之后)可知插入在2之后
>>> np.newaxis is None
True
二维情况:
>>> x = np.array([[1,2,3],[4,5,6]])
>>> print x
>>> print x.shape
[[1 2 3]
[4 5 6]]
(2, 3)#两行三列,因为第一个大括号下有两组list故是2,第一个大括号下的任意一组[]下又有3个数,故为3.
>>> y = np.expand_dims(x,axis=0)#axis=0表示最外面加[]
>>> print y
>>> print "y.shape: ",y.shape
>>> print "y[0][1]: ",y[0][1]
[[[1 2 3]
[4 5 6]]]
y.shape: (1, 2, 3)
y[0][1]: [4 5 6]
>>> y = np.expand_dims(x,axis=1)#在第二个[]下再加括号
>>> print y
>>> print "y.shape: ",y.shape
>>> print "y[1][0]: ",y[1][0]
[[[1 2 3]]
[[4 5 6]]]
y.shape: (2, 1, 3)
y[1][0]: [4 5 6]
>>> y = np.expand_dims(x,axis=2)#在第个括号里添加括号,以此作为新的一维
>>> print y
>>> print "y.shape: ",y.shape
[[[1]
[2]
[3]]
[[4]
[5]
[6]]]
y.shape: (2, 3, 1)
3.3、np.newaxis(新插入一个维度轴)
>>> a=np.array([1,2,3,4,5])
>>> b=a[np.newaxis,:]#插在最前面,从左到右,就按照从高维到低维位置排,在最面加个括号,增加了一维
>>> print a.shape,b.shape
>>> print a
>>> print b
(5,) (1, 5)
[1 2 3 4 5]
[[1 2 3 4 5]]
>>> a=np.array([1,2,3,4,5])
>>> b=a[:,np.newaxis]#插在第二个位置,从左到右,就按照从高维到低维位置排,在里面加个括号
>>> print a.shape,b.shape
>>> print a
>>> print b
(5,) (5, 1)
[1 2 3 4 5]
[[1]
[2]
[3]
[4]
[5]]
参考资料网址:
1、numpy 下的axis(轴)详细含义,np.expand_dims(x,axis=0),np.newaxis解释:https://blog.csdn.net/qq_35860352/article/details/80463111
2、np.concatenate的用法:https://www.jianshu.com/p/a094a954ff61