The difference between CONDA and pip

python Why is there a package management tool for pip We also need conda？ 

  I've been using it conda and pip , Sometimes the two are often mixed . But today, I found that the two things are not in the same place , So I found some things installed by myself , But when running the environment, I find that the package is always unrecognized , I've always been very confused , Until today, I have noticed this problem , So let's summarize the difference between the two .

Brief summary

pip

• pip Specialized management Python package
• Compile everything in the source code . （ Source code installation ）
• By the core Python Supported by the community （ namely ,Python 3.4+ Includes automatic enhancements pip Code for ）.

conda

• Python Agnostics . The main focus of the existing software package is Python, and conda In itself Python Compiling , But you can also be C Library or R The software package or any other software package provides conda software package .
• Install binaries . There is one named conda build Tools for , It can build packages from source code , but conda install It will install the built conda What's in the package .
• external . Conda yes Anaconda Package manager , from Continuum Analytics Provided Python Distribution version , But it can also be in Anaconda Use outside . You can use existing Python install , adopt pip Install it （ Although unless you have good reason to use an existing installation , Otherwise, this is not recommended ）.

Further said that ：

• conda It allows you to manage the installation process at the same time Relevant python Tasks and follow python Unrelated tasks , namely pip Can be installed in any environment python package ,conda Allow you to be in conda Install any language pack... In the environment （ Include C Language or python）.
• conda Use a new package format , You can't alternate conda and pip, because pip Cannot install and resolve conda Package format for . You can use these two tools , But they can't interact .
• because pip Is installed from source code , So if you can't compile the source code , Then installing it can be painful （ This is in Windows This is especially true in , But if the package has some difficulties C or FORTRAN library , Even in Linux It's the same with me ） rely on ）.
• Conda Install from binaries , That means someone （ for example Continuum） The hard work of compiling the software package has been completed , So the installation is very simple . however , Personal feeling conda Installation is a little slow .（ relative pip Come on ）
• If you are interested in building your own software package , There are also some differences . for example ,pip Based on the setuptools above , and conda Use its own format , There are some advantages to this （ For example, static , and Python I don't know ）.

original text ：《pip and conda The difference between 》