python numpy基础（一）基本用法

旧城等待， 2022-06-06 02:10 276阅读 0赞

（一）查看numpy版本

# 查看numpy版本
    import numpy as np
    print np.version.version
    # 1.11.3

（二）numpy多维数组为numpy.ndarray

# numpy多维数组为numpy.ndarray
    import numpy as np
    # 参数可以为list或tuple
    print np.array([1,2,3,4])
    # [1 2 3 4]
    print np.array([[1,2,3],[4,5,6]])
    # [[1 2 3]
    # [4 5 6]]
    print np.array((1.5,2,3,4))
    # [ 1.5 2. 3. 4. ]
    print np.array(((1,2,3),(4,5,6)))
    # [[1 2 3]
    # [4 5 6]]
    print type(np.array([1,2,3,4]))
    # <type 'numpy.ndarray'>

(三)numpy数据类型设定dtype和转换astype

# numpy数据类型设定dtype和转换astype
    import numpy as np
    array_str = np.array(['1.2','1.3','1.456'],dtype=np.string_)
    print array_str
    # ['1.2' '1.3' '1.456']
    # 从string类型转换为float类型
    print array_str.astype(float)
    # [ 1.2 1.3 1.456]

（四）numpy深拷贝和浅拷贝问题

# 关于numpy引用问题,因为numpy本身是拿来处理大数据的，所以x,y同时指向同一块内存空间
    import numpy as np
    x = np.array([1,2])
    y = x
    y[1] = 3
    print x
    # [1 3]
    
    # 如果想独立开来
    import numpy as np
    x = np.array([1,2])
    y = x.copy()
    y[1] = 3
    print x
    # [1 2]

（五）numpy多维数组转置，索引，切片

# numpy取唯一元素
    # np.unique(x)
    # numpy转置.T
    import numpy as np
    x = np.array([1,2]).T
    # numpy切片
    import numpy as np
    arr = np.arange(10)
    print arr
    # [0 1 2 3 4 5 6 7 8 9]
    arr[5:8] = 12
    print arr
    # [ 0 1 2 3 4 12 12 12 8 9]
    arr2 = arr[1:6]
    print arr2
    # [ 1 2 3 4 12]
    arr = np.arange(1,10).reshape(3,3)
    print arr
    # [[1 2 3]
    # [4 5 6]
    # [7 8 9]]
    print arr[1,2]
    # 6
    arr = arr.reshape(9,1)
    print arr
    # [[1]
    # [2]
    # [3]
    # [4]
    # [5]
    # [6]
    # [7]
    # [8]
    # [9]]
    print arr[5,0]
    # 6
    arr = arr.reshape(3,3)
    # 对数组内的各个元素进行比较
    print arr == 3
    # [[False False True]
    # [False False False]
    # [False False False]]
    # 如果数组中的值小于6则都为0
    arr[arr<6] = 0
    print arr
    # [[0 0 0]
    # [0 0 6]
    # [7 8 9]]

（六）numpy数组文件输出和输入,保存格式为.npy

import numpy as np
    arr = np.arange(10)
    np.save('D:\\machinetest\\array',arr)
    # 从文件加载数组
    arr = np.load('D:\\machinetest\\array.npy')
    print arr
    # [0 1 2 3 4 5 6 7 8 9]
    # 从txt读取数据loadtxt,savetxt为写入txt,默认读取数据为float型，需要指定数据类型dtype
    # 例如data1.txt:
    # aa bb cc
    # dd ee rr
    # 例如data2.txt
    # aa,bb,cc
    # dd,ee,rr
    arr1 = np.loadtxt('D:\\machinetest\\data1.txt',delimiter=' ',dtype=np.string_)
    print arr1
    arr2 = np.loadtxt('D:\\machinetest\\data2.txt',delimiter=',',dtype=np.string_)
    print arr2
    # [['aa' 'bb' 'cc']
    # ['dd' 'ee' 'rr']]

(七)使用numpy初始化数据

# 使用numpy初始化数据
    # numpy.arange
    import numpy as np
    arr = np.arange(10)
    print arr
    # [0 1 2 3 4 5 6 7 8 9]
    # 从1-5,平均生成9个值
    print np.linspace(1,5,9)
    # [ 1. 1.5 2. 2.5 3. 3.5 4. 4.5 5. ]
    # 初始化0
    np.zeros((3,4))
    # [[ 0. 0. 0. 0.]
    # [ 0. 0. 0. 0.]
    # [ 0. 0. 0. 0.]]
    # 初始化1
    np.ones((3,4))
    # [[ 1. 1. 1. 1.]
    # [ 1. 1. 1. 1.]
    # [ 1. 1. 1. 1.]]
    # 生成对角矩阵
    print np.eye(3)
    # [[ 1. 0. 0.]
    # [ 0. 1. 0.]
    # [ 0. 0. 1.]]
    # 生成网格矩阵
    arr1 = np.arange(1,5,1)
    # [1 2 3 4]
    arr2 = np.arange(5,10,1)
    # [5 6 7 8 9]
    arr3,arr4 = np.meshgrid(arr1,arr2)
    print arr3
    # [[1 2 3 4]
    # [1 2 3 4]
    # [1 2 3 4]
    # [1 2 3 4]
    # [1 2 3 4]]
    print arr4
    # [[5 5 5 5]
    # [6 6 6 6]
    # [7 7 7 7]
    # [8 8 8 8]
    # [9 9 9 9]]
    # 用于画图，填充坐标系,可使用以下方法
    # from matplotlib.colors import ListedColormap
    # m,n = arr4.shape
    # Z = np.eye(m)
    # print Z
    # cmap_light = ListedColormap(['#FFAAAA', '#AAFFAA'])
    # plt.pcolormesh(arr3,arr4,Z,cmap=cmap_light)

(八)获取数组的属性

# 获取数组的属性
    import numpy as np
    arr = np.zeros((2,3,4))
    print arr
    # [[[ 0. 0. 0. 0.]
    # [ 0. 0. 0. 0.]
    # [ 0. 0. 0. 0.]]
    # 
    # [[ 0. 0. 0. 0.]
    # [ 0. 0. 0. 0.]
    # [ 0. 0. 0. 0.]]]
    # 数组的维度
    print arr.ndim
    # 3
    # 数组每一维的大小
    print arr.shape
    # (2, 3, 4)
    # 数组的元素数
    print arr.size
    # 24
    # 数组里面元素的类型
    print arr.dtype
    # float64
    # 每个元素所占的字节数
    print arr.itemsize
    # 8

(九)numpy矩阵的计算

# numpy矩阵的计算
    import numpy as np
    # 全累加
    sum = np.sum([[0,1,2],[2,1,3]])
    print sum
    # 9
    # 往x轴压缩
    sum = np.sum([[0,1,2],[2,1,3]],axis=0)
    print sum
    # [2 2 5]
    # 往y轴压缩
    sum = np.sum([[0,1,2],[2,1,3]],axis=1)
    print sum
    # [3 6]
    # 往某个平面进行压缩
    arr = np.arange(27).reshape(3,3,3)
    print arr
    # [[[ 0 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]]]
    # 往xy平面进行压缩
    print np.sum(arr,axis=0)
    # [[27 30 33]
    # [36 39 42]
    # [45 48 51]]
    # 往xz平面进行压缩
    print np.sum(arr,axis=1)
    # [[ 9 12 15]
    # [36 39 42]
    # [63 66 69]]
    # 往yz平面进行压缩
    print np.sum(arr,axis=2)
    # [[ 9 12 15]
    # [36 39 42]
    # [63 66 69]]
    # 矩阵的转置arr.t==arr.transpose()
    arr = np.array([[1,0],[2,3]])
    print arr
    print arr.transpose()
    print arr.T
    
    #求均值
    import numpy as np
    a = np.array([[1, 2], [3, 4]])
    print(np.mean(a))
    # 2.5 全部相加后求平均
    print(np.mean(a, axis=0))
    # [ 2. 3.] 按维度来相加后求平均
    print(np.mean(a, axis=1))
    # [ 1.5 3.5]

（十）numpy数组的合并

# numpy数组的合并,进行的是深拷贝，即arr1和arr2合并生成的arr3，arr3和arr1,arr2互不影响
    import numpy as np
    arr1 = np.ones((2,2))
    arr2 = np.eye(2)
    # 垂直添加
    print np.vstack((arr1,arr2))
    # [[ 1. 1.]
    # [ 1. 1.]
    # [ 1. 0.]
    # [ 0. 1.]]
    # 水平添加
    print np.hstack((arr1,arr2))
    # [[ 1. 1. 1. 0.]
    # [ 1. 1. 0. 1.]]
    print np.c_[np.array([1,2,3]), np.array([4,5,6])]
    # [[1 4]
    # [2 5]
    # [3 6]]
    print np.c_[np.array([[1,2,3]]), 0, 0, np.array([[4,5,6]])]
    # [[1 2 3 0 0 4 5 6]]
    # 以上都需要数组知道是2列的，如果数组只有一列，则合并使用np.append
    print np.append([[1],[2],[3]],[[4],[5],[6]])
    # [1 2 3 4 5 6]