In the latest upgrade from Microsoft, its flagship database will have Python scripts running with complete access third party libraries in the form of native T-SQL procedures that are stored.
Then the process of computation will become much easier for the users. Unlike before, now the users will be able to perform their computations easily by using Python without any hassle.
With the Community Technology Preview SQL server 2017 release, it has become clear that Microsoft has considered the primary rule of information science. As result of this drastic change SQL server will be able to support Python to execute any kind of analytical work, also the data powered works, and execute machine learning models.
This modifications are also made available for the free-to-use Express edition, along with the SQL server 2017 Enterprise edition.
Generally the most commonly used technique in Python is to use SQL server as the data source with normal execution of Python scripts. It has now been made possible by Microsoft to insert Python code in SQL Server databases directly by incorporating it as a T-SQL stored procedure. The Python code can now be utilized in production with the same data that it will process.
If required both Python and R T-SQL codes can be used simultaneously in the same database. These features along with the RevoScalePy package were developed for SQL server by Microsoft when the R language was being integrated into the database.
It is not mandatory to have previous Python versions to use these new features. Also SQL server 2017 installs its own CPython edition (stock Python Interpreter accessible from the Python.org website) at the time of the set-up procedure. It also allows users to add their custom Python packages or create C code from various Python based modules by utilizing CPython to enjoy additional speeds.
Packages from the Python’s Anaconda distribution that are widely utilized in data science and Microsoft’s RevoScalePy package which contain a group of functions that analyze data which exploits certain features of SQL server such as column-store and in-memory index are included in this set up. The processing of SQL server with GPU-accelrated functions can be enhanced by third party modules like TensorFlow. In order to prevent violation of the network policies or security, constraints related to the Python runtime’s behavior can be set by the Database admins.
Whether other versions of Python will be allowed in CPython’s place is not clear, as some of these versions have been specifically created to increase productivity. An example would be the Python version distributed by Intel, a revamp of the Anaconda version, which utilizes the Intel Math Kernel Library to boost the processing of common mathematical operations like fast Fourier transforms or linear algebra on Intel processors. Excluding that, older math libraries like Pandas or NumPy for CPython are expected to provide the same type of enhancements they provide for Python in general also for Python in SQL Server.
Although Microsoft proclaimed earlier that Linux users will also be able to avail the benefits of using SQL Server, but as per the present scenario, only SQL Server 2017 (windows version) is Python supported.