Press "Enter" to skip to content

Category: T-SQL

TRY: CAST, CONVERT, and PARSE

Published 2023-12-26 by Kevin Feasel

Andy Brownsword tries and tries and tries again:

In the previous post we looked at how ISNUMERIC and TRY_CAST work and why we may want to utilise the latter when building validation for our data. When SQL Server 2012 rolled around it wasn’t only TRY_CAST which was added, we also had TRY_CONVERT and TRY_PARSE introduced too.

Here we’re going to look at how those function and the differences in the outputs which they can provide. We’ll start with the sample data below as the basis for these demonstrations:

Andy focuses mostly on the behavioral differences, where I like TRY_PARSE() a lot, especially because you can control locale for dates and times. When it comes to performance, I’ve found TRY_CAST() and TRY_CONVERT() to be essentially the same performance-wise. Maybe there’s a tiny difference between the two but there’s no guarantee of it.

TRY_PARSE(), however, calls into the .NET CLI for each execution and will be considerably slower. If you’re parsing a few or a few dozen values, you probably won’t notice the difference. If you’re parsing tens of thousands of values (or more), I assure you that you’ll notice the difference.

Comments closed

Advent of Code Day 6

Published 2023-12-22 by Kevin Feasel

Kevin Wilkie continues the advent of code series. The first part builds a small tally table and a loop:

Today we’re going racing! Sadly, it’s so not F1 or NASCAR racing. Snail racing is more like it since we’re moving millimeters by the end, but at least we’re closer to getting snow back to the elves, so let’s go racing!

Given a few numbers that are times and current record distances, this actually doesn’t look too bad to work with. First, as always, we have to load our data into SQL Server. This time, I loaded all of it into one table.

The second part goes back to the big tally table:

Sadly, this does make our numbers rather large, so we’re back to using the big ole Tally table we created for Walking Through Advent of Code Day 5.

This time I made it a little simpler on myself and just removed all of the spaces myself and placed the data in variables (one for time and one for distance). I thought this was an excellent idea since only one number would come out of all of this work.

Comments closed

Validating Numbers in T-SQL

Published 2023-12-21 by Kevin Feasel

Andy Brownsword asks if this thing is a number:

Data validation is key when ingesting from external sources. As we can’t always be certain of data quality we inevitably find bad data which needs to be handled. Here I wanted to look at a couple of options for validating numeric data.

Here’s the scenario – we’ve got data which may have been received via a flat file or passed into our database, and it should be a numeric value. How can we weed out the valid from invalid data?

Read on for the wrong answer (at least, the wrong answer given our expectations as developers or data platform specialists), followed by a good answer.

Comments closed

Advent of Code Day 5

Published 2023-12-20 by Kevin Feasel

Kevin Wilkie continues the advent of code. Part 1 starts with data prep and ends with some of the analysis:

There was a lot going on with this one, so let’s go ahead and get started. First thing we want to do, as always, is to go ahead and load all of our data into SQL Server. With this one, I cheated a little and loaded it into multiple tables to start with – since there are 7 different groups of data that we have to play with.

Could I have loaded it into one table and pull from there into other tables? Absolutely! Do I really want to? Heck, no!

Part 2 continues cross-referencing over a larger set of data:

This time, the meaning of the seed numbers is slightly changing on us. Instead of it only being 20 seeds we have to cross-reference (at least that’s how many seeds my input lists out), we now have 10 seeds and a rather large range of numbers to work with after each of those 10.

Thankfully, we can still use our Tally table to create yet another table with all of the seeds listed in it – yes, all of them!

Read on for Kevin’s solution.

Comments closed

Window Function Execution Plans with RANGE

Published 2023-12-20 by Kevin Feasel

Hugo Kornelis continues a series on explaining the execution plans for window functions:

This is part twenty-six of the plansplaining series. And already the fourth episode about window functions. The first of those posts covered basic window functions; the second post focused on fast-track optimization for running aggregates, and the third post explained how the optimizer works around the lack of execution plan support for UNBOUNDED FOLLOWING.

But all of those were about OVER specifications that use the ROWS keyword. Let’s now look at the alternative, the RANGE keyword.

Click through to see how the various options work with RANGE. By the way, I still want range intervals, like how Postgres implements them, where you can define an interval of X days/hours/minutes/whatever rather than a specific number of rows. Maybe one of these versions…

Comments closed

100 Estimated Rows

Published 2023-12-19 by Kevin Feasel

Deborah Melkin finds the optimizer in a “Dunno, here’s a guess” scenario:

A little while back, my coworker was having trouble with a query and asked me about a strange thing they were seeing in the execution plan: The estimated rows for the query was showing 100.

100 felt like an awfully specific number. And there were two scenarios I knew of where the SQL Server cardinality estimator immediately used that value – table variables and multi-statement table value functions with SQL 2016 compatibility or earlier or as part of an APPLY where interleaved execution doesn’t apply instead of a JOIN.

The statement in question didn’t use either so what was the issue?

Read on for the solution.

Comments closed

Searching for a String in a SQL Server Database

Published 2023-12-15 by Kevin Feasel

Vlad Drumea is looking for a string:

This post contains a script that I’ve written and used a few times when I’ve needed to search for a string in an entire SQL Server database.

You can find more of my scripts by checking out the SQL Server Scripts category of my blog.

As a warning, this is definitely going to be slow on larger databases. But for smaller databases, it can definitely be handy.

Comments closed

Advent of Code Day 4

Published 2023-12-14 by Kevin Feasel

Kevin Wilkie continues a Quixotic quest. Part 1 sets up the story:

Now we get to work with a set of numbers that our elf played and a set of winning numbers for each card. This sounds like something that SQL is meant for!

As always, we load the fun input data provided by the AoC group into our database. I actually loaded it in as 2 columns – Card and String. I’m just not a fan of throwing all of our data into one column of a table and letting it all get sorted out later.

Part 2 wraps it up:

With part 1, we just had to figure out how many times each of the winners showed up. With part 2 though, we have to jump through several hoops since we now have to determine how many cards we will end up with if we win the next series of cards for every match.

I slimmed down the table that I’m working with to only 2 columns – yes, in the real world, I would have used a view, but today was not that day.

By the way, as soon as I saw OverlyLongNamesThatNoOneCanTypeWithoutUsingIntellisense I wondered when Kevin got access to my code base. I have, on a few occasions, created punishment names, names so long that they punished the people who had to type or track them. In fairness, it wasn’t just a fit of pique, though pique was a common factor in all of those situations.

Comments closed

Batch Changes in T-SQL

Published 2023-12-14 by Kevin Feasel

Erik Darling doesn’t update all the rows at once like some barbarian:

The goal of batching modifications is to accomplish large numbers of row modifications with exacerbating locking problems, and being kinder to your server’s transaction log.

There is generally an appreciable difference in transaction time in modifying 1,000 rows and modifying 10,000,000 rows. Go try it, if you don’t believe me.

Click through for more information on good ways to set up batching, including the use of the highly underrated OUTPUT ... INTO ... clause on DML statements.

Comments closed