Category: T-SQL

Itzik Ben-Gan knows how to inspire joy:

Microsoft recently released the first public preview of SQL Server 2022. This release has a number of T-SQL improvements. In this article I focus on windowing and NULL-related improvements. These include the new WINDOW clause and the windowing NULL treatment clause.

I’ll be using the sample database TSQLV6 in the examples in this article. You can download this sample database here.

I’ve been waiting for INGORE NULLS since they previewed it in Azure SQL Edge.

Comments closed

Aaron Bertrand notes some T-SQL improvements:

A few of the most useful changes I’ve been able to play with in SQL Server 2022 so far:

– GREATEST / LEAST

– STRING_SPLIT

– DATE_BUCKET

– GENERATE_SERIES

In this tip, I’ll explain each one, and show some practical use cases.

Click through for more information. I particularly like GREATEST() and LEAST() but GENERATE_SERIES() could be very useful as well. Some of this stuff was first made available in Azure SQL Edge.

Comments closed

Chad Callihan shows off a good use of aggregate functions and the CASE statement:

Have you have been tasked with pulling multiple counts from the same table? Maybe you need to find how many records have a value for a column and how many are NULL. Or maybe you need to see how many records are true and how many are false.

It’s simple enough to run a query to count one set of criteria, run another query for the second set of criteria, and combine them when sending your results. Did you know you can get multiple counts with one query?

Here’s an example of how using COUNT and CASE can speed up your day.

Click through for an example. I mildly disagree with Chad’s conclusion that this is something you’ll rarely do—the more you work with reporting and analytical queries, the more you’ll appreciate this.

Comments closed

Daniel Hutmacher mixes separators with regular characters:

Suppose you have a delimited string input that you want to split into its parts. That’s what STRING_SPLIT() does:

DECLARE @source nvarchar(max)='Canada, Cape Verde, '+    'Central African Republic, Chad, Chile, China, Colombia, Comoros';

SELECT TRIM([value]) AS[Country]
FROM STRING_SPLIT(@source, ',');

Simple enough. But delimited lists are tricky, because the delimiter could exist in the name itself. Look for yourself what happens when we add the two Congos to the list:

Daniel has a clever solution to the problem.

Comments closed

Steve Jones shows off a good use case for window functions:

Recently I was looking at some data and wanted to analyze it by month. I have a goal that is set for each day and then an actual value. I wanted to know how I was tracking against the goal, as a running total. If my goal is 10 a day, then I ought to actually get to 10 the first day, 20 for the second day (10 + 10), etc.

Read on to see how Steve solved the problem.

Comments closed