Reducing the Number of Reads in Your Queries

“A-Bridged” – reducing reads – get it?  I’m sorry you’re reading this.  Photo by Jamie Street on Unsplash

In last week’s post, we went over how one of best ways to improve query performance was to reduce the number of reads that your query has to do.

Less reads typically means faster query performance – so how can you reduce the number of reads SQL Server is required to make?

Prefer a visual?  Watch this week’s content on my YouTube channel.

Write more selective queries

Writing more selective queries can be done in a  few different ways.

For starters, if you don’t need every column of data, don’t use SELECT *.  Depending on the size of your rows, providing only the columns you need might allow SQL Server to use an index  that is narrower and/or denser.  An index that is narrow (fewer columns) or dense (more records per page) allows SQL Server to return the same amount of data you need in fewer pages.

The same thing goes for being more selective in your ON and WHERE clauses.  If you specify no WHERE conditions, SQL Server will return every single page in your table.  If you can filter down the data to exactly what you need, SQL Server can return only the data you need. This reduces logical reads and improves speed.

Finally, do you have large object (LOB) data (eg. varchar(max), image, etc..) on the tables in your query? Do you actually need it in your final result set? No?  Then don’t include it as part of your reads!  This could mean creating an index on the columns you do need or putting your LOB data in a separate table and only joining to it when you need it.

Fix suboptimal execution plans

It’s possible that your query is already as selective as it can be.  Maybe you are getting too many reads because SQL Server is generating and using a suboptimal execution plan.

A big tip off that this might be happening is if your cardinality estimates are out of whack a.k.a. the estimated vs. actual row counts have a large difference between them:

If SQL Server thinks it only is going to read 1 row of data, but instead needs to read way more rows of data, it might choose a poor execution plan which results in more reads.

You might get a suboptimal execution plan like above for a variety of reasons, but here are the most common ones I see:

If you had a query that previously ran fine but doesn’t anymore, you might be able to utilize Query Store to help identify why SQL Server started generating suboptimal plans.

The key is to get a good execution plan so that you aren’t performing unnecessary reads.

But what if you aren’t encountering any of the problems above and performance is still slow due to high numbers of reads?  Simply…

Add an index

A cup of coffee and a shot of espresso might have the same caffeine content – espresso is just more caffeine dense, just like the data stored in a narrow index.  Photo by Mike Marquez on Unsplash

If you have an existing table that has many columns and you only need a subset of them for your query, then consider adding an index for those columns.

Indexes are copies of your data stored in a different order with generally fewer columns.  If SQL Server is able to get all of the information it needs from a narrow index, it will do that instead of reading the full table/clustered index.

A copy of the data that has fewer columns will have greater page density (or the amount of data that fits on each page).  If SQL Server can get all of the data it needs by reading fewer, denser pages then your query will run faster.

Don’t just go adding indexes willy nilly though.  You may already have an index that almost contains all of the columns your query needs.  Look at a table’s existing indexes first and see if any of them are close to what you need.  Usually, you’ll be better off adding an included column or two to an existing index instead creating a whole brand new index.  And you’ll save on disk space by not creating duplicate indexes either.

Reduce index fragmentation

So indexes are great for reducing reads because they allow us to store only the data that is needed for a specific query (both as key columns and included columns).  Fewer columns = greater density = fewer reads necessary.

However, indexes can become  fragmented.  There’s internal fragmentation, which causes less data to be stored on a page than what is possible, and external fragmentation which causes the pages to be stored out of logical order on the disk.

Internal fragmentation is problematic because it reduces page density, causing SQL Server to have to read more pages in order to get all of the data it needs.

External fragmentation is problematic, especially for spinning disk hard drives, because SQL Server needs to read from all over the disk to get the data it needs.

In general, reorganizing or rebuilding an index are the typical ways you want to fix a fragmented index.

To slow down future fragmentation, you can test out different fill factors to try and prevent page splits from fragmenting your indexes.

Thanks for reading. You might also enjoy following me on Twitter.

How to use Statistics IO to Improve Your Query Performance

SQL Server’s STATISTICS IO reporting is a great tool to help you performance tune queries.
Usually the goal of performance tuning is to make your query run faster.  One of the easiest ways to get a faster query is to reduce the amount of data a query is processing.  STATISTICS IO makes it easy to see how much data SQL Server is actually processing.
Specifically, the STATISTICS IO output helps with performance tuning because:
  1. The data it shows acts as a measuring stick for your performance tuning changes.
  2. It provides a good way of isolating the query changes you are making from other changes that may be happening on the server.

Rather watch than read?  Head over to my YouTube channel and watch this week’s content instead.

So what’s STATISTICS IO data look like?

To show IO statistics on your query, you first need to execute:

After running a query with with the above setting turned on, check SQL Server Management Studio’s Messages tab to see output that looks something like this:

The key things that my eyes are drawn to on initial examination of the STATISTICS IO output are the following:

  • Logical reads: The number of 8kB pages SQL Server had to read from the buffer cache (memory) in order to process and return the results of your query.  The more pages that need to be read, the slower your query.
  • Worktables/Workfiles: These are temporary objects that SQL Server creates in tempdb in order to process query results.  Although not always bad, it might indicate that SQL is doing more work than it needs to (perhaps an index could help?)
  • Lob Logical Reads: The number of large objects (e.g. varchar(max)) SQL is having to read.  I take the most cursory glance at this – if I’m returning high numbers of lobs, I might want to make sure I actually need them.  If not, I may add an index or move the lobs off to a separate table.

There are more properties in the STATISTICS IO output, but if we can significantly decrease the above three indicators then chances are good that we’ll improve our query performance.

So why are these three indicators so useful?

Tracking performance changes

The main reason I like performance tuning with STATISTICS IO is because it makes it easy to create a baseline for my query performance.  When I make changes to the query, it’s then easy to see if my changes helped or hurt the query.

The main metric I use for this is logical reads.  Logical reads refers to pages pulled from the cache (memory) versus physical reads which indicates the number of pages from disk.  However, all pages get loaded from disk into cache before SQL Server is able to use them.

This makes logical reads great for tracking performance changes because it clearly tells me how many 8kB pages in total SQL Server needed to read in order to my return my data.

If I add an index, does the total number of pages read go up or down?  Let me check my logical read counts and see.

What if I add some additional filtering or restructure my query?  I can easily tell if my changes hurt performance by seeing if the total number of logical reads went up or down.

Logical reads allow me to easily track the effectiveness of my tuning tactics.

The same concept applies to my worktable and logical lob read properties.  For the former, any time SQL Server is having to write data out to disk (tempdb in this case), performance will be slower.

In the latter case, if SQL Server is needing to move around large objects that comprise of multiple 8kB pages each, things will be slow.  If I can keep track of how many lob logical reads SQL Server is performing, then I can focus on removing that overhead from my query.

Isolating other factors that impact performance

For the same reason logical reads make it easy to track query performance after making changes, using logical reads makes it easy to mute other factors that might affect performance.

For example, I used to think that simply watching how long it took a query to run was a good indicator of helping me performance tune.  If the total number of seconds it took my query to run decreased, my changes helped improve performance!

This is a potentially deceptive way to measure performance though because what if during my first run the server was getting slammed by other queries?  Using elapsed run time isn’t an effective way to measure performance.

Also, server environment hardware isn’t always the same.  I might test a query in one environment and then deploy it to another.  My testing on an empty dev box might have been great, but as soon as the query runs in production along with all other queries, it might not perform as well.

More than one way to analyze a query

STATISTICS IO is a great place to start your performance tuning process.

It doesn’t mean that you have to stop there though.  While being able to track the effects of your tuning changes and isolating other environment variables is important, ultimately you will have to use other means to actually improve performance.

So be sure to look at execution plans, dig into index and table statistics, rewrite the order of your table joins to see if it makes a difference, etc…  Just remember, performance always comes back to how much data SQL Server needs to process – reduce that and your queries will surely perform better.

Thanks for reading. You might also enjoy following me on Twitter.

Does The Join Order of My Tables Matter?

Photo by pan xiaozhen on Unsplash

I had a great question submitted to me (thank you Brandman!) that I thought would make for a good blog post:

…I’ve been wondering if it really matters from a performance standpoint where I start my queries. For example, if I join from A-B-C, would I be better off starting at table B and then going to A & C?

The short answer: Yes.  And no.

More of a watcher than a reader?  Watch this week’s episode on YouTube!

Table join order matters for performance!

Disclaimer: For this post, I’m only going to be talking about INNER joins.  OUTER (LEFT, RIGHT, FULL, etc…) joins are a whole ‘nother animal that I’ll save for time.

Let’s use the following query from WideWorldImporters for our examples:

Note: with an INNER join, I normally would prefer putting my ‘USA’ filter in the WHERE clause, but for the rest of these examples it’ll be easier to have it part of the ON.

The key thing to notice is that we are joining  three tables – Orders, OrderLines, and StockItems – and that OrderLines is what we use to join between the other two tables.

We basically have two options for table join orders then – we can join Orders with OrderLines first and then join in StockItems, or we can join OrderLines and StockItems first and then join in Orders.

In terms of performance, it’s almost certain that the latter scenario (joining OrderLines with StockItems first) will be faster because StockItems will help us be more selective.

Selective?  Well you might notice that our StockItems table is small with only 227 rows.  It’s made even smaller by filtering on ‘USA’ which reduces it to only 8 rows.

Since the StockItems table has no duplicate rows (it’s a simple lookup table for product information) it is a great table to join with as early as possible since it will reduce the total number of rows getting passed around for the remainder of the query.

If we tried doing the Orders to OrderLines join first, we actually wouldn’t filter out any rows in our first step, cause our subsequent join to StockItems to be more slower (because more rows would have to be processed).

Basically, join order DOES matter because if we can join two tables that will reduce the number of rows needed to be processed by subsequent steps, then our performance will improve.

So if the order that our tables are joined in makes a big difference for performance reasons, SQL Server follows the join order we define right?

SQL Server doesn’t let you choose the join order

SQL is a declarative language: you write code that specifies *what* data to get, not *how* to get it.

Basically, the SQL Server query optimizer takes your SQL query and decides on its own how it thinks it should get the data.

It does this by using precalculated statistics on your table sizes and data contents in order to be able to pick a “good enough” plan quickly.

So even if we rearrange the order of the tables in our FROM statement like this:

Or if we add parentheses:

Or even if we rewrite the tables into subqueries:

SQL Server will interpret and optimize our three separate queries (plus the original one from the top of the page) into the same exact execution plan:

Basically, no matter how we try to redefine the order of our tables in the FROM statement, SQL Server will still do what it thinks it’s best.

But what if SQL Server doesn’t know best?

The majority of the time I see SQL Server doing something inefficient with an execution plan it’s usually due to something wrong with statistics for that table/index.

Statistics are also a whole ‘nother topic for a whole ‘nother day (or month) of blog posts, so to not get too side tracked with this post, I’ll point you to Kimberly Tripp’s introductory blog post on the subject: https://www.sqlskills.com/blogs/kimberly/the-accidental-dba-day-15-of-30-statistics-maintenance/

The key thing to take away is that if SQL Server is generating an execution plan where the order of table joins doesn’t make sense check your statistics first because they are the root cause of many performance problems!

Forcing a join order

So you already checked to see if your statistics are the problem and exhausted all possibilities on that front.  SQL Server isn’t optimizing for the optimal table join order, so what can you do?

Row goals

If SQL Server isn’t behaving and I need to force a table join order, my preferred way is to do it via a TOP() command.

I learned this technique from watching Adam Machanic’s fantastic presentation on the subject and I highly recommend you watch it.

Since in our example query SQL Server is already joining the tables in the most efficient order, let’s force an inefficient join by joining Orders with OrderLines first.

Basically, we write a subquery around the tables we want to join together first and make sure to include a TOP clause. 

Including TOP forces SQL to perform the join between Orders and OrderLines first – inefficient in this example, but a great success in being able to control what SQL Server does.

This is my favorite way of forcing a join order because we get to inject control over the join order of two specific tables in this case (Orders and OrderLines) but SQL Server will still use its own judgement in how any remaining tables should be joined.

While forcing a join order is generally a bad idea (what happens if the underlying data changes in the future and your forced join no longer is the best option), in certain scenarios where its required the TOP technique will cause the least amount of performance problems (since SQL still gets to decide what happens with the rest of the tables).

The same can’t be said if using hints…

Query and join hints

Query and join hints will successfully force the order of the table joins in your query, however they have significant draw backs.

Let’s look at the FORCE ORDER query hint.  Adding it to your query will successfully force the table joins to occur in the order that they are listed:

Looking at the execution plan we can see that Orders and OrderLines were joined together first as expected:

The biggest drawback with the FORCE ORDER hint is that all tables in your query are going to have their join order forced (not evident in this example…but imagine we were joining 4 or 5 tables in total).

This makes your query incredibly fragile; if the underlying data changes in the future, you could be forcing multiple inefficient join orders.  Your query that you tuned with FORCE ORDER could go from running in seconds to minutes or hours.

The same problem exists with using a join hints:

Using the LOOP hint successfully forces our join order again, but once again the join order of all of our tables becomes fixed:

A join hint is probably the most fragile hint that forces table join order because not only is it forcing the join order, but it’s also forcing the algorithm used to perform the join.

In general, I only use query hints to force table join order as a temporary fix.

Maybe production has a problem and I need to get things running again; a query or join hint may be the quickest way to fix the immediate issue.  However, long term using the hint is probably a bad idea, so after the immediate fires are put out I will go back and try to determine the root cause of the performance problem.

Summary

  • Table join order matters for reducing the number of rows that the rest of the query needs to process.
  • By default SQL Server gives you no control over the join order – it uses statistics and the query optimizer to pick what it thinks is a good join order.
  • Most of the time, the query optimizer does a great job at picking efficient join orders.  When it doesn’t, the first thing I do is check to see the health of my statistics and figure out if it’s picking a sub-optimal plan because of that.
  • If I am in a special scenario and I truly do need to force a join order, I’ll use the TOP clause to force a join order since it only forces the order of a single join.
  • In an emergency “production-servers-are-on-fire” scenario, I might use a query or join hint to immediately fix a performance issue and go back to implement a better solution once things calm down.

Thanks for reading. You might also enjoy following me on Twitter.