在SQL Server 2017中使用Python进行数据插值和转换

As a continuation to my previous article, How to use Python in SQL Server 2017 to obtain advanced data analytics, a little bit of curiosity about Deep Learning with Python integration in SQL Server led me to write this latest article.

作为我上一篇文章如何在SQL Server 2017中使用Python来获取高级数据分析的续篇， 对SQL Server中 Python集成的深度学习的好奇心使我撰写了这篇最新文章。

With Python running within SQL Server, you can bring the existing data and the code together. Data is accessible directly, so there’s no need to extract query data sets, moving data from storage to the application. It’s a useful approach, especially considering issues of data sovereignty and compliance, since the code runs within the SQL Server security boundaries, triggered by a single call from T-SQL stored procedures.

通过在SQL Server中运行Python ，您可以将现有数据和代码结合在一起。 数据可直接访问，因此无需提取查询数据集，即可将数据从存储区移至应用程序。 这是一种有用的方法，尤其是考虑到数据主权和合规性问题时，因为代码是在T-SQL存储过程的一次调用的触发下在SQL Server安全边界内运行的。

As we’re all aware, Microsoft is now taking some steps that had us surprised, one of them being the release of SQL Server 2017 to Linux. In an attempt to expand the horizons, Microsoft has brought in Python capabilities within SQL Server. Python is a language that is easily learned and it packs a lot of potentials. There’s no better time to learn Python, since enterprises are already changing gears to use IT to better derive value from their businesses data.

众所周知，Microsoft现在正在采取一些令我们感到惊讶的步骤，其中之一就是将SQL Server 2017发行到Linux。 为了扩大视野，Microsoft在SQL Server中引入了Python功能。 Python是一种易于学习的语言，它具有许多潜力。 没有更好的时间来学习Python，因为企业已经在改变使用IT手段以更好地从其业务数据中获取价值的手段。

Machine learning and artificial intelligence (AI) may sound intimidating but actually, it’s a great value-add to organizations in the areas such as web search, financial prediction, fraud detection, digitization etc. There is a lot of hype around deep learning and AI, and the question, whether to use it or not to use it but one thing we can all agree upon is that analytics is of a lot of value to businesses.

机器学习和人工智能（AI）听起来可能令人生畏，但实际上，它对组织在网络搜索，财务预测，欺诈检测，数字化等领域具有巨大的附加值。深度学习和AI周围有很多炒作，还有一个问题，即是否使用它，但我们都可以同意的一件事是，分析对企业具有很大的价值。

It’s really hard to say at the initial stage how well this integration with SQL Server would be or how well SQL Server can withstand the data science design and methodology, but it sure is an interesting feature and a highly rated data science product to be tested in SQL Server 2017.

在初始阶段，真的很难说与SQL Server的集成程度如何，或者SQL Server可以承受数据科学设计和方法的程度如何，但是它确实是一个有趣的功能，并且需要在其中进行高度评价的数据科学产品进行测试。 SQL Server 2017。

The available options made me execute several Python samples and create a few data mining examples to understand the core concepts of data analytics. As I mentioned in my previous article How to use Python in SQL Server 2017 to obtain advanced data analytics, it’s all about data loading and data transformation. If the model is huge, one may have a hard time loading the data and transferring it over the network. Hence, data scientists do their predictive analysis using the sampling method.

可用的选项使我可以执行多个Python示例，并创建一些数据挖掘示例以了解数据分析的核心概念。 正如我在上一篇文章如何在SQL Server 2017中使用Python来获取高级数据分析所提到的，所有这些都与数据加载和数据转换有关。 如果模型很大，则可能很难加载数据并通过网络传输。 因此，数据科学家使用采样方法进行预测分析。

If you are a database administrator interested in leveraging data science, your first question would be, “Where do I start and how?”

如果您是对利用数据科学感兴趣的数据库管理员，那么您的第一个问题将是：“我从哪里开始？如何开始？”

从哪里开始 (Where to begin)

As a database administrator, the thought o