Microsoft SQL Server 2017’s support for Python had me curious. Would this integration make it practical to use Python in non-data analytic scenarios where a general-purpose programming language—like Python—is more suitable than set-based T-SQL?

The answer to this question hinges around the integration’s design. Presumably, the new Python integration is designed to work well for data analysis, as this seems to be the prime impetus motivating it. An integration that excels in one scenario, like the assemble data -> process & analyze -> return results flow common in data analytics, may or may not be a good fit for other use cases.

To help us evaluate when and where SQL Server’s Python support may be helpful for non-data analytics applications, let’s compare it—or, more specifically, let’s compare the underlying external script execution environment that powers it—with two integrations that have been included with SQL Server for some time: Common Language Runtime (CLR) and xp_cmdshell.
Continue reading →