Catching circular references in parent-child structures

A popular form of organizing dimensions is in parent-child structures, also known as “unbalanced” or “ragged” dimensions, because any branch can have an arbitrary number of child levels. There are many advantages to this type of representation, but their recursive nature also brings some challenges. In this post, we’re going to look at circular references, and how you can trap them before they run out of control.

Different query plans for “OR” type queries

The SQL Server query optimizer can find interesting ways to tackle seemingly simple operations that can be hard to optimize. Consider the following query on a table with two indexes, one on (a), the other on (b):

SELECT a, b
FROM #data
WHERE a<=10 OR b<=10000;

The basic problem is that we would really want to use both indexes in a single query.

In this post, we’re going to take a look at a few examples of how this type of query would be optimized, as well as how statistics can affect the query plan, and finally, we’ll take a look at a slightly rare plan operator called “Merge Join (Concatenation)”.

Intermittent conversion issues

Although it’s rotten practice, values like dates, times, numbers, etc are often stored in columns with the wrong data type. Sometimes, it’s because the data model needs to allow for multiple types of data in the same column (sometimes even on the same row in the form of a formula or text syntax), but sometimes, it’s just plain and simple lazy.

With this type of setup, it’s only a matter of time before you run into conversion errors like this one:

Msg 245, Level 16, State 1, Line 29
Conversion failed when converting the varchar value 'xyz' to data type int.

To make things worse, these errors can be tricky to pinpoint, because they can appear to come and go without any real recognizable pattern. In this post, I’ll take a look at how this happens, how to look for these errors, and ultimately how to fix them if you can’t change the database schema.

Median and percentile in T-SQL

Back in 2014 I wrote a blog post on how to calculate a median value using the NTILE window function. It’s far from the best performing solution there is, but it worked on SQL Server 2008, before the introduction of OFFSET-FETCH i SQL Server. In this post, I’m going to look at creating a generalized function that calculates the median (or any percentile) of a series of values.

PIVOT, UNPIVOT and performance

Few things deserve the attention of a long rant as much as unneccessarily complicated syntaxes. When you want to achieve something that is clearly defined and supported, but you have to look up the the syntax. PIVOT and UNPIVOT are examples of such features, and in this case, I’ll even show you a more well-performing alternative.

The 2015 Swedish SQL Server usergroup challenge

It’s been absolutely ages since I last wrote a blog post, mostly because I’ve been busy getting my shiny new own consultancy up to speed, but I’ll admit that writer’s block has also been a factor.

But here’s something to write home about. This year’s annual Swedish SQL Server usergroup challenge was as interesting as ever, and it marks my third stab at this prestigious competition. In this post, I’ll go through my contribution, highlighting some of the techniques that I’ve applied to make it go really fast.

Joining two SCD2 tables

A number of OLTP systems store dimension data in SCD2-like tables in order to retain all the revisions whenever the dimension information changes. In certain situations, you may come across a need to join two or more SCD tables, while keeping all the versions information intact. Sound tricky? Not really.

Distributing values with overlapping ranges

Today, we’re going to look at how to distribute values with what I call “overlapping ranges”. What I mean by that is that the value that you need to distribute has a given range and the objects you want to distribute the values to are also defined as ranges. Sound complicated? It doesn’t have to.

Manual transaction management

Transactions are great. They keep your data together atomically, so you’re not in for any nasty surprises. But even a novice knows better than to leave transactions open, waiting for user interaction. If you do, lock waits and probably deadlocks will pile up in no time.

So how do you book a flight without blocking all the other users or losing your seat to somebody else while you make up your mind?