Scribd
Upload
8 views

NumPy_and_NumPy_Arrays

Uploaded by

gotip76351
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
8 views

NumPy_and_NumPy_Arrays

Uploaded by

gotip76351
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

NumPy and NumPy Arrays

What is NumPy?

NumPy, which stands for Numerical Python, is a powerful library for


numerical and mathematical operations in Python. It provides support
for large, multi-dimensional arrays and matrices, along with a collection
of high-level mathematical functions to operate on these arrays.

Originally, Numeric, crafted by Jim Hugunin, served as the precursor to


NumPy. Another package, Numarray, was concurrently developed,
offering additional functionalities. In 2005, Travis Oliphant integrated the
features of Numarray into the Numeric package, giving rise to the
creation of the NumPy package. Numerous contributors have played a
role in advancing this open-source.

Data analysis, enhanced by tools like NumPy, involves inspecting,


cleaning, transforming, and modeling data. NumPy facilitates efficient
numerical operations and array handling, aiding in discovering insights
and supporting decision-making processes.

Operations with NumPy: With NumPy, developers can execute the


following operations:
• Mathematical and logical operations on arrays.
• Fourier transforms and functions for manipulating shapes.
• Operations associated with linear algebra. NumPy incorporates
built-in functions for both linear algebra and random number
generation. Numpy

Setting Up NumPy Environment:

The regular Python setup doesn't include NumPy. A simple solution is to


add NumPy to your Python using the popular installer called pip.

pip install numpy

1|Page
You can check if NumPy is installed successfully by opening a Python
interpreter and trying to import NumPy:

import numpy as np

If there are no error messages, NumPy is successfully installed.

To effectively activate NumPy, it is recommended to install a binary


package tailored for your operating system. These packages
encompass the complete SciPy stack, incorporating NumPy, SciPy,
matplotlib, IPython, SymPy, and nose packages, in addition to the core
Python components.

For Windows-

• Anaconda, accessible at continuum.io, provides a complimentary


Python distribution encompassing the complete SciPy stack, and it
is compatible with Linux and Mac operating systems.
• Canopy, accessible at enthought.com/products/canopy/, offers
both a free and a commercial distribution that includes the entire
SciPy stack, catering to Windows, Linux, and Mac users.
• Python(x, y), available for free, is a Python distribution equipped
with the SciPy stack and the Spyder IDE, designed specifically for
the Windows operating system. It can be downloaded from
python-xy.github.io/.

Linux users can employ the package managers of their respective


distributions to install individual or multiple packages within the SciPy
stack.

Numpy Arrays: It is an N-dimensional array designed to hold a


collection of items of the same data type. Accessing these items is
achieved using a zero-based indexing scheme. It is also called ndarray.
Here are key characteristics and concepts related to NumPy arrays:

1. Homogeneous Data Types: NumPy arrays have a fixed data type,


which ensures that all elements in the array are of the same type.
This allows for efficient storage and computation.

2|Page
2. Multidimensional: NumPy arrays can have multiple dimensions
(axes). For example, a 1-dimensional array is like a list, a 2-
dimensional array is like a matrix, and a 3-dimensional array is like
a cube of numbers.

3. Zero-Based Indexing: NumPy arrays use a zero-based indexing


scheme, which means the index of the first element is 0, the
second element is at index 1, and so on.

4. Broadcasting: NumPy supports broadcasting, a powerful


mechanism that allows operations between arrays of different
shapes and sizes. Broadcasting automatically adjusts the
dimensions of smaller arrays to perform element-wise operations
with larger arrays.

5. Vectorized Operations: NumPy supports vectorized operations,


which means that you can perform operations on entire arrays
without the need for explicit loops.

6. Memory Efficiency and Data Type Specification: Each element


within a ndarray occupies a fixed-size memory block, and its data
type is specified by the dtype attribute. When extracting a portion
of a ndarray using slicing, the resulting object belongs to one of
the array scalar types.

The dtype defines the kind of data that can be stored in the ndarray,
such as integers, floats, or strings. The header stores information about
the ndarray, such as its shape and size.

3|Page
NumPy arrays, represented by the ndarray object, form the backbone of
numerical computing in Python. They offer a convenient and efficient
way to work with large datasets, enabling a wide range of mathematical
and scientific operations.

Numpy Datatypes: NumPy offers a broader range of numerical types


compared to Python. The table below illustrates some common data
types defined in NumPy.

1. bool_: Represents Boolean values (True or False) stored as a byte.


2. int_: Default integer type, akin to C long; typically, int64 or int32.
3. intc: Corresponds to C int, usually int32 or int64.
4. int8: Represents a byte, ranging from -128 to 127.
5. int16: Represents a 16-bit integer, ranging from -32768 to 32767.
6. int32: Represents a 32-bit integer, ranging from -2147483648 to
2147483647.
7. int64: Represents a 64-bit integer, ranging from -
9223372036854775808 to 9223372036854775807.
8. uint8: Represents an unsigned 8-bit integer, ranging from 0 to 255.
9. float_: Shorthand for float64, representing double precision
floating-point numbers.
10. float16: Represents a half-precision float with a sign bit, 5 bits
exponent, and 10 bits mantissa.
11. float32: Represents a single-precision float with a sign bit, 8 bits
exponent, and 23 bits mantissa.
12. float64: Represents a double-precision float with a sign bit, 11 bits
exponent, and 52 bits mantissa.
13. complex_: Shorthand for complex128, representing double
precision complex numbers

Data Type Objects (dtype): Data type Objects in NumPy define how to
interpret a fixed block of memory corresponding to an array. The
interpretation depends on several factors:

4|Page
1. Type of data: Specifies whether the data is an integer, float, or a
Python object.

2. Size of data: Defines the size of the data in terms of bytes.

3. Byte order: Determines the byte order, whether it is little-endian or


big-endian.

4. For structured types: Includes the names of fields, the data type of
each field, and the portion of the memory block occupied by each
field.

5. For subarrays: Specifies the shape and data type of the subarray.

Syntax: Construct a dtype object

numpy.dtype(object, align, copy)

The parameters are:

• Object: The entity to be converted into a dtype object.

• Align: If true, it adds padding to the field to align it with a C-struct.

• Copy: If set to true, it creates a new copy of the dtype object. If false,
the result is a reference to the built-in data type object.

Example Code:

# using array-scalar type


import numpy as np
dt = np.dtype(np.int32)
print(dt)

Output
int32

The code defines a NumPy data type (int32) using the array-scalar type
constructor and prints the data type.

5|Page

You might also like