Abstract:

The increasing demand for real-time analytics and scalable statistical processing has driven the development of in-database analytical frameworks that integrate statistical methods directly into database platforms. SAP HANA Predictive Analysis Library (PAL) exemplifies this approach by providing a collection of statistical and machine learning algorithms executed within an in-memory database environment. This article presents a concise overview of selected statistical techniques available in PAL, including the Chi-Square Goodness-of-Fit test, the Chi-Square Test of Independence, and the Cumulative Distribution Function (CDF).
The paper outlines the theoretical foundations of these methods and demonstrates their practical implementation using in-database analytics. The Chi-Square Goodness-of-Fit test is discussed as a tool for evaluating conformity between observed and expected distributions, while the Chi-Square Test of Independence is presented as a method for analyzing relationships between categorical variables. Additionally, the role of CDF in probabilistic analysis and distribution characterization is examined. Practical examples with SQL and Python (hana-ml) illustrate how statistical analysis can be performed directly within SAP HANA, reducing data movement and improving performance. The study highlights the advantages of embedded analytics for scalable enterprise data analysis.