[883] Install Python package using Jupyter Notebook

In the installation of packages in Jupyter Notebook, I encountered an issue. I installed packages using Command Prompt (CMD), but when I try to import them in Jupyter Notebook, it prompted the error that "No module named XXX". This blog is trying to solve this issue.

ref: Install Python package using Jupyter Notebook

ref: 解决python Jupyter不能导入外部包问题

---

Installing Python Library in Jupyter

Using ! pip install

To install Python libraries, we use pip command on the command line console of the Operating System. The OS has a set of paths to executable programs in its so-called environment variables through which it identifies directly what exactly the pip means. This is the reason that whenever pip command can directly be run on the console.

In Jupyter, the console commands can be executed by the ‘!’ sign before the command within the cell. For example, If the following code is written in the Jupyter cell, it will execute as a command in CMD.

! echo GeeksforGeeks

Similarly we can install any package via jupyter in the same way, and it will run it directly in the OS shell.

Syntax:

! pip install [package_name] 

But using this method is not recommended because of the OS behavior. This command executed on the current version in the $PATH variable of the OS. So in the case of multiple Python versions, this might not install the same package in the jupyter’s Python version. In the simplest case it may work.

Using sys library

To solve the above-mentioned problem, it is recommended to use sys library in Python which will return the path of the current version’s pip on which the jupyter is running. sys.executable will return the path of the Python.exe of the version on which the current Jupyter instance is

Syntax:

import sys
!{sys.executable} -m pip install [package_name]

By the above code, the package will be installed in the same Python version on which the jupyter notebook is running.