Python for the Oracle DBA – Connects to everything

In my introduction to my Python for the Oracle DBA topic I said that Python can connect to everything that an Oracle DBA needs. I want to use this post to expand on that and talk about the Python modules and packages that I have used.

As you might expect, Python includes built-in modules and ones that you can add. Python comes with a large built-in library called the Python Standard Library. To use these modules you just import them into your scripts. There is nothing to install. You can also install new modules from PyPI, the Python Package Index. You install a PyPI package using a tool called pip that acts a lot like Red Hat Linux’s yum utility. Here is the pip command line help:

D:\>pip

Usage:
  pip  [options]

Commands:
  install                     Install packages.
  download                    Download packages.
  uninstall                   Uninstall packages.
 ... etc. ...

The plotting module that I use, Matplotlib,  is part of a larger group of modules called SciPi which I assume stands for Scientific Python. I have only used Matplotlib and Numpy but SciPi includes a number of other modules. You can get the various SciPy modules from PyPI using pip as described above.

Obviously an Oracle DBA wants to connect to Oracle databases. The cx_Oracle package lets you connect to an Oracle database. I am still using version 5.3 but a newer version is out and supports a variety of features. cx_Oracle requires an Oracle client. I have been mainly using a 32 bit 11.2 Oracle Windows client with cx_Oracle 5.3 on my Windows 7 corporate laptop. There is a Python standard for database APIs so if you install modules to connect to various types of databases they will hopefully follow this standard and have similar interfaces. I have connected to Snowflake, MySQL, and Big Query databases using Python as well but have not done anything significant with them.

I have written a couple of scripts on Linux that I would have written using bash shell scripting in the past. I describe in an earlier post how you can use the built-in subprocess module to run SQL*Plus from Python. More recently I wrote a script to test all of my Unix passwords with the Paramiko package which does ssh. Here is the script:

It just tries to log into a host using ssh and prints out a useful message if the login fails. It runs the hostname command after it logs in. Note the import statements at the top of the program. I installed the paramiko module from PyPI but sys and socket are part of the standard library.

The last thing I want to mention is how you can use Python with Amazon Web Services or AWS. I have started to learn how to work with AWS for my job and have done some very basic Python scripts to just connect to AWS and run simple commands. Amazon calls their main Python module Boto 3. We have production Python scripts that other people in my company have written so Python is alive and well in Amazon’s cloud. It probably would not hurt for an Oracle DBA to learn some Python just because we might need it when working with cloud services. I have already mentioned the cloud based databases Snowflake and Big Query so those are also examples of Python working with the cloud.

So, I have mainly given an overview of the Python packages and modules that I have used. I have only played with some of them. I have used others such as cx_Oracle, Matplotlib, Paramiko, and subprocess for real work. Based on my experience so far I think Python can connect to just about anything I would need as an Oracle DBA.

Bobby