Python-NumPy

╰+攻爆jí腚メ 2022-06-04 07:56 202阅读 0赞

#### 网上查到的个人认为最简洁的大纲，如果你有Python基础，又有些数学基础，看这个就够了。看不懂的地方敲一遍，这篇文章能告诉我们numpy是什么、它能做什么、怎么做。 ####

# NumPy #

NumPy是高性能科学计算和数据分析的基础包。部分功能如下：

*   *  ndarray, 具有矢量算术运算和复杂广播能力的快速且节省空间的多维数组。
     *  用于对整组数据进行快速运算的标准数学函数（无需编写循环）。
     *  用于读写磁盘数据的工具以及用于操作内存映射文件的工具。
     *  线性代数、随机数生成以及傅里叶变换功能。
     *  用于集成C、C++、Fortran等语言编写的代码的工具。

首先要导入numpy库：import numpy as np

#### 　　A NumPy函数和属性： ####

表2.1.A.1 NumPy类型

<table style="margin:0px; padding:0px; border:1px solid silver; border-collapse:collapse; word-break:break-word"> 
 <tbody style="margin:0px; padding:0px"> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 生成函数</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 作用</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.array( x) np.array( x, dtype) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 将输入数据转化为一个ndarray 将输入数据转化为一个类型为type的ndarray </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.asarray( array )</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 将输入数据转化为一个新的（copy）ndarray</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.ones( N ) np.ones( N, dtype) np.ones_like( ndarray ) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 生成一个N长度的一维全一ndarray 生成一个N长度类型是dtype的一维全一ndarray 生成一个形状与参数相同的全一ndarray </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.zeros( N) np.zeros( N, dtype) np.zeros_like(ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 生成一个N长度的一维全零ndarray 生成一个N长度类型位dtype的一维全零ndarray 类似np.ones_like( ndarray ) </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.empty( N ) np.empty( N, dtype) np.empty(ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 生成一个N长度的未初始化一维ndarray 生成一个N长度类型是dtype的未初始化一维ndarray 类似np.ones_like( ndarray ) </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.eye( N ) np.identity( N ) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 创建一个N * N的单位矩阵（对角线为1，其余为0） </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.arange( num) np.arange( begin, end) np.arange( begin, end, step) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 生成一个从0到num-1步数为1的一维ndarray 生成一个从begin到end-1步数为1的一维ndarray 生成一个从begin到end-step的步数为step的一维ndarray </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.mershgrid(ndarray, ndarray,...) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 生成一个ndarray * ndarray * ...的多维ndarray </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.where(cond, ndarray1, ndarray2) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 根据条件cond，选取ndarray1或者ndarray2，返回一个新的ndarray </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.in1d(ndarray, [x,y,...]) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 检查ndarray中的元素是否等于[x,y,...]中的一个，返回bool数组 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> &nbsp;</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> &nbsp;</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 矩阵函数</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 说明</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.diag( ndarray) np.diag( [x,y,...]) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 以一维数组的形式返回方阵的对角线（或非对角线）元素 将一维数组转化为方阵（非对角线元素为0） </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.dot(ndarray, ndarray)</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 矩阵乘法</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.trace( ndarray)</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 计算对角线元素的和</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> &nbsp;</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> &nbsp;</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> &nbsp;</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> &nbsp;</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 排序函数 </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 说明 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.sort( ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 排序，返回副本 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.unique(ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 返回ndarray中的元素，排除重复元素之后，并进行排序 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.intersect1d( ndarray1, ndarray2) np.union1d( ndarray1, ndarray2) np.setdiff1d( ndarray1, ndarray2) np.setxor1d( ndarray1, ndarray2) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 返回二者的交集并排序。 返回二者的并集并排序。 返回二者的差。 返回二者的对称差 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> &nbsp;</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> &nbsp;</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 一元计算函数</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 说明</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.abs(ndarray) np.fabs(ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 计算绝对值 计算绝对值（非复数） </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.mean(ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 求平均值 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.sqrt(ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 计算x^0.5 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.square(ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 计算x^2 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.exp(ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 计算e^x </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> log、log10、log2、log1p </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 计算自然对数、底为10的log、底为2的log、底为(1+x)的log </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.sign(ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 计算正负号：1（正）、0（0）、-1（负） </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.ceil(ndarray) np.floor(ndarray) np.rint(ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 计算大于等于改值的最小整数 计算小于等于该值的最大整数 四舍五入到最近的整数，保留dtype </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.modf(ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 将数组的小数和整数部分以两个独立的数组方式返回 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.isnan(ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 返回一个判断是否是NaN的bool型数组 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.isfinite(ndarray) np.isinf(ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 返回一个判断是否是有穷（非inf，非NaN）的bool型数组 返回一个判断是否是无穷的bool型数组 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> cos、cosh、sin、sinh、tan、tanh </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 普通型和双曲型三角函数 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> arccos、arccosh、arcsin、arcsinh、arctan、arctanh </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 反三角函数和双曲型反三角函数 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.logical_not(ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 计算各元素not x的真值，相当于-ndarray </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 多元计算函数 </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 说明 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.add(ndarray, ndarray) np.subtract(ndarray, ndarray) np.multiply(ndarray, ndarray) np.divide(ndarray, ndarray) np.floor_divide(ndarray, ndarray) np.power(ndarray, ndarray) np.mod(ndarray, ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 相加 相减 乘法 除法 圆整除法（丢弃余数） 次方 求模 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.maximum(ndarray, ndarray) np.fmax(ndarray, ndarray) np.minimun(ndarray, ndarray) np.fmin(ndarray, ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 求最大值 求最大值（忽略NaN） 求最小值 求最小值（忽略NaN） </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.copysign(ndarray, ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 将参数2中的符号赋予参数1 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.greater(ndarray, ndarray) np.greater_equal(ndarray, ndarray) np.less(ndarray, ndarray) np.less_equal(ndarray, ndarray) np.equal(ndarray, ndarray) np.not_equal(ndarray, ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> &gt; &gt;= &lt; &lt;= == != </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> logical_and(ndarray, ndarray) logical_or(ndarray, ndarray) logical_xor(ndarray, ndarray) </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> &amp; | ^ </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.dot( ndarray, ndarray)</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 计算两个ndarray的矩阵内积</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.ix_([x,y,m,n],...)</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 生成一个索引器，用于Fancy indexing(花式索引)</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> &nbsp;</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> &nbsp;</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 文件读写</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 说明</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.save(string, ndarray)</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 将ndarray保存到文件名为 [string].npy 的文件中（无压缩）</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.savez(string, ndarray1, ndarray2, ...)</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 将所有的ndarray压缩保存到文件名为[string].npy的文件中</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.savetxt(sring, ndarray, fmt, newline=‘\n‘)</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 将ndarray写入文件，格式为fmt</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.load(string)</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 读取文件名string的文件内容并转化为ndarray对象（或字典对象）</td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> np.loadtxt(string, delimiter)</td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 读取文件名string的文件内容，以delimiter为分隔符转化为ndarray</td> 
 </tr> 
 </tbody> 
</table>

表2.1.A.2 np常用函数

#### 　　B NumPy.ndarray函数和属性： ####

表2.1.B.1 ndarray属性

表2.1.B.2 ndarray函数

表2.1.B.3 ndarray索引/切片方式

#### 　　C NumPy.random函数和属性： ####

2.1.C.1 random常用函数

#### 　　D NumPy.linalg函数和属性： ####