Creating a large database

SQL Data Aggregation Aggravation

When we have to deal with and store a lot of data, it makes sense to aggregate it so that we store only the information we actually need. If we get this right, this works well, but the design of the system takes care and thought because the problems can be subtle and various. Joe Celko describes some of the ways that things can go wrong and end up providing incorrect, inaccurate or misleading results.

Temporary tables SCOPE

Statistics in SQL: The Mann–Whitney U test

Phil Factor shows how to use the Mann-Whitney U test in SQL to to find out whether two samples come from the same distribution.

The Order of Processing

Saving For Christmas

Statistics in SQL: The Kruskal–Wallis Test

Before you report your conclusions about your data, have you checked whether your 'actionable' figures occurred by chance? The Kruskal-Wallis test is a safe way of determining whether samples come from the same population, because it is simple and doesn't rely on a normal distribution in the population. This allows you a measure of confidence that your results are 'significant'. Phil Factor explains how to do it.

The Basics of Good T-SQL Coding Style – Part 2: Defining Database Objects

Technical debt is a real problem in database development, where corners have been cut in the rush to keep to dates. The result may work but the problems are in the details: such things as inconsistent naming of objects, or of defining columns; sloppy use of data types, archaic syntax or obsolete system functions. With databases, technical debt is even harder to pay back. Robert Sheldon explains how and why you can get it right first time instead.

SQL Server User-Defined Functions

User-Defined Functions (UDFs) are an essential part of the database developers' armoury. They are extraordinarily versatile, but just because you can even use scalar UDFs in WHERE clauses, computed columns and check constraints doesn't mean that you should. Multi-statement UDFs come at a cost and it is good to understand all the restrictions and potential drawbacks. Phil Factor gives an overview of User-defined functions: their virtues, vices and their syntax.

Simple SQL: Attribute Splitting

If the design of a relational database is wrong, no amount of clever DML SQL will make it work well. Dr. Codd’s Information Principle is that you have, inside the entity tables, the columns that model the attributes of that entity. The columns contain scalar values. Tables that model relationships can have attributes, but they must have references to entities in the schema. You split those attributes at your peril. Joe Celko explains the basics.