Anaconda vs. miniconda

In the Anaconda repository, there are two types of installers:

"Anaconda installers" and "Miniconda installers".

What are their differences? Besides, for an installer file, Anaconda2-4.4.0.1-Linux-ppc64le.sh, what does 2-4.4.0.1 stand for?

Answers


The difference is that miniconda is just shipping the repository management system. So when you install it there is just the management system without packages. Whereas with Anaconda, it is like a distribution with some built in packages.

Like with any Linux distribution, there are some releases which bundles lots of updates for the included packages. That is why there is a difference in version numbering. If you only decide to upgrade Anaconda, you are updating a whole system.


Per the original docs (link is now dead):

Choose Anaconda if you:

  • Are new to conda or Python
  • Like the convenience of having Python and over 150 scientific packages automatically installed at once
  • Have the time and disk space (a few minutes and 3 GB), and/or
  • Don’t want to install each of the packages you want to use individually.

Choose Miniconda if you:

  • Do not mind installing each of the packages you want to use individually.
  • Do not have time or disk space to install over 150 packages at once, and/or
  • Just want fast access to Python and the conda commands, and wish to sort out the other programs later.

I use Miniconda myself. Anaconda is bloated. Many of the packages are never used and could still be easily installed if and when needed.

Note that Conda is the package manager (e.g. conda list displays all installed packages in the environment), whereas Anaconda and Miniconda are distributions. A software distribution is a collection of packages, pre-built and pre-configured, that can be installed and used on a system. A package manager is a tool that automates the process of installing, updating, and removing packages.

Anaconda is a full distribution of the central software in the PyData ecosystem, and includes Python itself along with the binaries for several hundred third-party open-source projects. Miniconda is essentially an installer for an empty conda environment, containing only Conda, its dependencies, and Python. Source.

Once Conda is installed, you can then install whatever package you need from scratch along with any desired version of Python.

2-4.4.0.1 is the version number for your Anaconda installation package. Strangely, it is not listed in their Old Package Lists.

In April 2016, the Anaconda versioning jumped from 2.5 to 4.0 in order to avoid confusion with Python versions 2 & 3. Version 4.0 included the Anaconda Navigator.

Release notes for subsequent versions can be found here.


Miniconda gives you the Python interpreter itself, along with a command-line tool called conda which operates as a cross-platform package manager geared toward Python packages, similar in spirit to the apt or yum tools that Linux users might be familiar with.

Anaconda includes both Python and conda, and additionally bundles a suite of other pre-installed packages geared toward scientific computing. Because of the size of this bundle, expect the installation to consume several gigabytes of disk space.

Source: Jake VanderPlas's Python Data Science Handbook


The 2 in Anaconda2 means that the main version of Python will be 2.x rather than the 3.x installed in Anaconda3. The current release has Python 2.7.13.

The 4.4.0.1 is the version number of Anaconda. The current advertised version is 4.4.0 and I assume the .1 is a minor release or for other similar use. The Windows releases, which I use, just say 4.4.0 in the file name.

Others have now explained the difference between Anaconda and Miniconda, so I'll skip that.


Both Anaconda and miniconda use the conda package manager. The chief differece between between Anaconda and miniconda,however,is that

The Anaconda distribution comes pre-loaded with all the packages while the miniconda distribution is just the management system without any pre-loaded packages. If one uses miniconda, one has to download individual packages and libraries separately.

I personally use Anaconda distribution as I dont really have to worry much about individual package installations.

A disadvantage of miniconda is that installing each individual package can take a long amount of time. Compared to that installing and using Anaconda takes a lot less time.

However, there are some packages in anaconda (QtConsole, Glueviz,Orange3) that I have never had to use. I dont even know their purpose. So a disadvantage of anaconda is that it occupies more space than needed.


Anaconda is a very large installation ~ 2 GB and is most useful for those users who are not familiar with installing modules or packages with other package managers.

Anaconda seems to be promoting itself as the official package manager of Jupyter. It's not. Anaconda bundles Jupyter, R, python, and many packages with its installation.

Anaconda is not necessary for installing Jupyter Lab or the R kernel. There is plenty of information available elsewhere for installing Jupyter Lab or Notebooks. There is also plenty of information elsewhere for installing R studio. The following shows how to install the R kernel directly from R Studio:

To install the R kernel, without Anaconda, start R Studio. In the R terminal window enter these three commands:

install.packages("devtools")
devtools::install_github("IRkernel/IRkernel")
IRkernel::installspec()

Done. The next time Jupyter is opened, the R kernel will be available and available.


Brief

conda is both a command line tool, and a python package.

Miniconda installer = Python + conda

Anaconda installer = Python + conda + meta package anaconda

meta Python pkg anaconda = about 160 Python pkgs for daily use in data science

Anaconda installer = Miniconda installer + conda install anaconda

Detail
  1. conda is a python manager and an environment manager, which makes it possible to

    • install package with conda install flake8
    • create an environment with any version of Python with conda create -n myenv python=3.6
  2. Miniconda installer = Python + conda

    conda, the package manager and environment manager, is a Python package. So Python is installed. Cause conda distribute Python interpreter with its own libraries/dependencies but not the existing ones on your operating system, other minimal dependencies like openssl, ncurses, sqlite, etc are installed as well.

    Basically, Miniconda is just conda and its minimal dependencies. And the environment where conda is installed is the "base" environment, which is previously called "root" environment.

  3. Anaconda installer = Python + conda + meta package anaconda

  4. meta Python package anaconda = about 160 Python pkgs for daily use in data science

    Meta packages, are packages that do NOT contain actual softwares and simply depend on other packages to be installed.

    Download an anaconda meta package from Anaconda Cloud and extract the content from it. The actual 160+ packages to be installed are listed in info/recipe/meta.yaml.

    package:
        name: anaconda
        version: '2019.07'
    build:
        ignore_run_exports:
            - '*'
        number: '0'
        pin_depends: strict
        string: py36_0
    requirements:
        build:
            - python 3.6.8 haf84260_0
        is_meta_pkg:
            - true
        run:
            - alabaster 0.7.12 py36_0
            - anaconda-client 1.7.2 py36_0
            - anaconda-project 0.8.3 py_0
            # ...
            - beautifulsoup4 4.7.1 py36_1
            # ...
            - curl 7.65.2 ha441bb4_0
            # ...
            - hdf5 1.10.4 hfa1e0ec_0
            # ...
            - ipykernel 5.1.1 py36h39e3cac_0
            - ipython 7.6.1 py36h39e3cac_0
            - ipython_genutils 0.2.0 py36h241746c_0
            - ipywidgets 7.5.0 py_0
            # ...
            - jupyter 1.0.0 py36_7
            - jupyter_client 5.3.1 py_0
            - jupyter_console 6.0.0 py36_0
            - jupyter_core 4.5.0 py_0
            - jupyterlab 1.0.2 py36hf63ae98_0
            - jupyterlab_server 1.0.0 py_0
            # ...
            - matplotlib 3.1.0 py36h54f8f79_0
            # ...
            - mkl 2019.4 233
            - mkl-service 2.0.2 py36h1de35cc_0
            - mkl_fft 1.0.12 py36h5e564d8_0
            - mkl_random 1.0.2 py36h27c97d8_0
            # ...
            - nltk 3.4.4 py36_0
            # ...
            - numpy 1.16.4 py36hacdab7b_0
            - numpy-base 1.16.4 py36h6575580_0
            - numpydoc 0.9.1 py_0
            # ...
            - pandas 0.24.2 py36h0a44026_0
            - pandoc 2.2.3.2 0
            # ...
            - pillow 6.1.0 py36hb68e598_0
            # ...
            - pyqt 5.9.2 py36h655552a_2
            # ...
            - qt 5.9.7 h468cd18_1
            - qtawesome 0.5.7 py36_1
            - qtconsole 4.5.1 py_0
            - qtpy 1.8.0 py_0
            # ...
            - requests 2.22.0 py36_0
            # ...
            - sphinx 2.1.2 py_0
            - sphinxcontrib 1.0 py36_1
            - sphinxcontrib-applehelp 1.0.1 py_0
            - sphinxcontrib-devhelp 1.0.1 py_0
            - sphinxcontrib-htmlhelp 1.0.2 py_0
            - sphinxcontrib-jsmath 1.0.1 py_0
            - sphinxcontrib-qthelp 1.0.2 py_0
            - sphinxcontrib-serializinghtml 1.1.3 py_0
            - sphinxcontrib-websupport 1.1.2 py_0
            - spyder 3.3.6 py36_0
            - spyder-kernels 0.5.1 py36_0
            # ...
    

    The pre-installed packages from meta pkg anaconda are mainly for web scraping and data science. Like requests, beautifulsoup, numpy, nltk, etc.

    If you have a Miniconda installed, conda install anaconda will make it same as an Anaconda installation, except that the installation folder names are different.

  5. Miniconda2 v.s. Miniconda. Anaconda2 v.s. Anaconda.

    2 means the bundled Python interpreter for conda in the "base" environment is Python 2, but not Python 3.


Need Your Help

JNI vs. JNA performance

java c concurrency java-native-interface jna

We have a native c/asm application utilizing GPU(OpenCL) for a big encrypt/decrypt data with a specific method, and it just works perfect, no problem. A part of the project (web and distribution) i...

change app name in App Store

ios xcode itunesconnect

I have app called "X". It's already approved in the app store. Now I want it to be called "X Pro" for some reasons. What should I edit in plist or ITC to change this name? I guess I should change