Category: Internals

Row Goals On Anti-Joins

Published 2018-03-09 by Kevin Feasel

Paul White continues his row goals series:

The optimizer assumes that people write a semi join (indirectly e.g. using EXISTS) with the expectation that the row being searched for will be found. An apply semi join row goal is set by the optimizer to help find that expected matching row quickly.

For anti join (expressed e.g. using NOT EXISTS) the optimizer’s assumption is that a matching row will not be found. An apply anti join row goal is not set by the optimizer, because it expects to have to check all rows to confirm there is no match.

If there does turn out to be a matching row, the apply anti join might take longer to locate this row than it would if a row goal had been used. Nevertheless, the anti join will still terminate its search as soon as the (unexpected) match is encountered.

This is a shorter article but very useful in understanding row goals, along with the rest of his series.

Comments closed

Using Extended Properties For Documentation

Published 2018-03-05 by Kevin Feasel

Phil Factor shows us how we can use Extended Properties to build database documentation:

Once you’ve got into the habit of using Extended Properties to document your database, there are obvious benefits:

You can explain why you added that index or modified that constraint.

You can describe exactly what that rather obscure column does.

You can add a reasoned explanation to the use of a table.

You will often need these explanations because, sadly, DDL code isn’t ‘self-documenting’, and human memory is fallible. Extended Properties are easily searched because they are all exposed in one system view.

It is great to add explanations to lists of procedures, functions and views once the database becomes sizeable. Extended Properties are useful when exploring the metadata, but the requirement isn’t quite so essential because comments are preserved along with the source code. Tables, however, are a big problem because SQL Server throws away the script that produces the table, along with all the comments. The reason that this happens is that there are many ways you can alter parts of a table without scripting the entire table. How could one infallibly preserve all these ALTER statements in the preserved script? It’s tricky. Table scripts that you get from SSMS or script via SMO are therefore synthesised from the system tables but without those comments or even Extended Properties.

Extended Properties are useful, but I think the lack of tooling around them prevented widespread adoption. Now that there are a few tools which support them (including SSMS’s data classification mechanism), I wonder if these will get a second look.

Comments closed

Row Goals And Semi Joins

Published 2018-02-23 by Kevin Feasel

Paul White continues his row goals series:

The remaining physical join type is nested loops, which comes in two flavours: regular (uncorrelated) nested loops and apply nested loops (sometimes also referred to as a correlated or lateral join).

Regular nested loops join is similar to hash and merge join in that the join predicate is evaluated at the join. As before, this means there is no value in setting a row goal on either input. The left (upper) input will always be fully consumed eventually, and the inner input has no way to determine which row(s) should be prioritized, since we cannot know if a row will join or not until the predicate is tested at the join.

By contrast, an apply nested loops join has one or more outer references (correlated parameters) at the join, with the join predicate pushed down the inner (lower) side of the join. This creates an opportunity for the useful application of a row goal. Recall that a semi join only requires us to check for the existence of a row on join input B that matches the current row on join input A (thinking just about nested loops join strategies now).

In other words, on each iteration of an apply, we can stop looking at input B as soon as the first match is found, using the pushed-down join predicate. This is exactly the sort of thing a row goal is good for: generating part of a plan optimized to return the first n matching rows quickly (where n = 1 here).

This has the depth and quality that you naturally expect from Paul, making it an immediate read.

Comments closed

Digging Into The In-Memory Columnstore Location

Published 2018-02-12 by Kevin Feasel

Niko Neugebauer does some investigation into where, exactly, memory-optimized columnstore data goes:

This is a rather simple blog post that is dedicated to the theme of the In-Memory Columnstore Indexes location. This has been a constant topic of discussion over a long period of time, even during the public events – and there is a need to clear out this topic.

I have assumed that the In-Memory Columnstore structures (Segments, Dictionaries, …) are located in the In-Memory, but there have been voices that I greatly respect, pointing that actually the Columnstore Object Pool is the exact location of any Columnstore structures, and there is nothing better than to take this feature for a ride and see what the SQL Server engine is actually doing.

Niko shows off a couple of useful DMVs along the way, too.

Comments closed

The Links That Tie Row To LOB

Published 2018-02-07 by Kevin Feasel

Steve Stedman shows how to use DBCC PAGE and DBCC IND to piece together where LOB data is stored for a particular row:

The question came up as how to find a link from blog storage that is corrupt back to the table and row that contains that data.

The is no link from the blob storage back to the table and row, but this is a link from the data page containing the table and row off to the blob data.

Read the whole thing.

Comments closed

SQL Server Internal Row Structures

Published 2018-01-16 by Kevin Feasel

David Fowler gets to the guts of a row as stored in SQL Server:

DBCC page will take in a database name or id, file id and page id and return a representation of the specified page depending on the print options that you choose.

We’ve got four different print options that we can choose,

0 – Return only the page header
1 – Return the page header and hex dump of each row
2 – Return the page header and full page hex dump
3 – Return the page header, hex dump of each row as well as the details on each column

Read the whole thing.

Comments closed

SQL Server’s Referential Integrity Operator

Published 2017-12-06 by Kevin Feasel

Joe Obbish explains the purpose of the referential integrity operator in SQL Server 2016:

What would happen if a parent table was referenced by hundreds of child tables, such as for a date dimension table? Deleting or updating a row in the parent table would create a query plan with at least one join per incoming foreign key reference. Creating a query plan for that statement is equivalent to creating a query plan for a query containing hundreds or even thousands of joins. That query plan could take a long time to compile or could even time out. For example, I created a simple query with 2500 joins and it still hadn’t finished compiling after 15 minutes. That’s why I assume a table is limited to 253 incoming foreign key references in SQL Server 2014.

That restriction won’t be hit often but could be pretty inconvenient to work around. The referential integrity operator introduced with compatibility level 130 raises the limit from 253 to 10000. All of the joins are collapsed into a single operator which can reduce compile time and avoid errors.

There’s some really good information in this post, and Joe has mixed feelings on the concept.

Comments closed

How Non-Clustered Index Key Columns Are Stored

Published 2017-11-30 by Kevin Feasel

Kendra Little walks through page-level details on a non-clustered index:

Just like in the root page and the intermediate pages, the FirstName and RowID columns are present.

Also in the leaf: CharCol, our included column appears! It was not in any of the other levels we inspected, because included columns only exist in the leaf of a nonclustered index.

Kendra does a great job of explaining the topic.

Comments closed

Base Versus Simple Containment

Published 2017-11-28 by Kevin Feasel

Joe Obbish takes a crack at explaining the difference between base containment and simple containment for cardinality estimation:

We know that the first query will return 500k rows and the second query will return 0 rows. However, can SQL Server know that? Each statistics object only contains information about its own column. There’s no correlation between the UNIQUE_ID and MOD_FILTER columns, so there isn’t a way for SQL Server to know that the queries will return different estimates. The query optimizer can create an estimate based on the filters on the WHERE clause and on the histograms of the join columns, but there’s no foolproof way to do that calculation. The presence of the filters introduces uncertainty into the estimate, even with statistics that perfectly describe the data for each column. The containment assumption is all about the modeling assumption that SQL Server has to make to resolve that uncertainty.

It’s an interesting post aimed at trying to get you to think like a simplified cardinality estimator. SQL Server doesn’t behave exactly like this, but it’s a good mental reference point.

Comments closed

R Internals: Data Sizes With Nullable Columns

Published 2017-11-27 by Kevin Feasel

Niels Berglund digs into the Binary Exchange Langage (BXL) and notices something weird about data sizes:

When looking at the data sent, the size of the packages and “drilling” into the TCP packets we could deduct that: :

Each column has an over-head of 32 bytes (at least for non nullable data)
The size of the column in one row is the size of the data type for numeric types.
For decimal and numeric an extra byte is added to each column, where this byte indicates the precision.
Columns of alpha numeric type all had 2 bytes pre-pended to the bytes, except max types.
For char and nchar the storage size was 2 bytes plus the size the column was defined as.
For varchar and nvarchar the storage size was 2 bytes plus the size of the data stored.
For the varmax data types the number of bytes that were pre-pended varied dependent on the data size.

Read the whole thing.

Comments closed