Discovering New System Objects and Functions in SQL Server 2019

Problem

SQL Server 2019 CTP was recently released. There is some information posted about new features such as this document by Microsoft What is new in SQL Server 2019 and this article on mssqlTips What's New in the First Public CTP of SQL Server 2019 . What other new features are there that can be useful to a DBA's daily operations?

Solution

As a SQL Server DBA, we are always excited about a new release of SQL Server. I once wrote a tip Identify System Object Differences Between SQL Server Versions and by using the script in that tip, I have explored the new objects and new changes for existing objects, such as new columns for views/tables or new parameters for functions/stored procedures.

Here are the detailed steps and some interesting findings in SQL Server 2019.

Step 1 Environment Setup

To find what’s new in SQL Server 2019, we need two versions of SQL Server. I will use SQL Server 2017 and SQL Server 2019. To follow along, you should have these two SQL Server instances installed.

Step 2 Collect Data Follow the detailed instructions in this previous tip, Identify System Object Differences Between SQL Server Versions , once this is done, we should have three inventory tables populated. I put all these three tables in [TempDB] for convenience. The three inventory tables will be shown in the analysis scripts below. Step 3 Analyze Data

Once step 2 is done, we can start to do some interesting exploration. We first take a look at the system object changes in SQL Server 2019 since SQL Server 2017.

-- 1. find new system objects
declare @tgt_ver int = 15; -- sql server 2019
declare @src_ver int = 14; -- sql server 2017
select [schema], [object], [type] from tempdb.dbo.allobject
where [version]<a href="/cdn-cgi/l/email-protection" data-cfemail="88b5c8fceffc">[email protected]</a>_ver
except
select [schema], [object], [type] from tempdb.dbo.allobject
where [version]<a href="/cdn-cgi/l/email-protection" data-cfemail="fbc6bb888998">[email protected]</a>_ver
go

We get 53 new objects, the last few are shown below:

-- 2. find dropped system objects
declare @tgt_ver int = 15; -- sql server 2019
declare @src_ver int = 14; -- sql server 2017
select [schema], [object], [type] from tempdb.dbo.allobject
where [version]<a href="/cdn-cgi/l/email-protection" data-cfemail="a994e9dadbca">[email protected]</a>_ver
except
select [schema], [object], [type] from tempdb.dbo.allobject
where [version]<a href="/cdn-cgi/l/email-protection" data-cfemail="221f62564556">[email protected]</a>_ver
go

The result is two extended stored procedures are dropped:

Now we take a look at the new columns added to system tables and views:

--3. new columns to sys views/tables
declare @tgt_ver int = 15; -- sql server 2019
declare @src_ver int = 14; -- sql server 2017
; with c as (
select distinct o.id, o.[schema], o.[object], o.[type], o.version--, p2.[object_type]
from dbo.AllObject o
inner join dbo.AllObject o2
on o.[schema]=o2.[schema]
and o.[object]=o2.[object]
and o.[version]<a href="/cdn-cgi/l/email-protection" data-cfemail="2e136e5a495a">[email protected]</a>_ver
and o2.[version]<a href="/cdn-cgi/l/email-protection" data-cfemail="2914695a5b4a">[email protected]</a>_ver
), w as (
select [Object]=ac.[schema]+'.'+ac.[Object]
, ac.[column], ac.column_type, ac.max_length, ac.[version]--, c.[version]
from dbo.AllColumn ac
inner join c
on c.[object]=ac.[object] and c.[schema]=ac.[schema] and c.[version]=ac.[version]
left join dbo.AllColumn ac2
on ac2.[schema]=ac.[schema] and ac2.object = ac.object
and ac2.[column]=ac.[column]
and ac2.[version]<a href="/cdn-cgi/l/email-protection" data-cfemail="526f12212031">[email protected]</a>_ver
where ac2.[column] is null)
select [object], [column], [column_type], max_length
from w
order by 1

We see 75 changes; a sample of the changes is listed as below:

Now we will check the parameter changes to functions and stored procedures:

--4. find new parameters added to SPs/functions
declare @tgt_ver int = 15;
declare @src_ver int = 14;
; with c as (
select distinct p1.id, p1.[schema], p1.[object], p1.[object_type]--, p2.[object_type]
from dbo.AllParam p1
inner join dbo.AllParam p2
on p1.[schema]=p2.[schema]
and p1.[object]=p2.[object]
and p1.[version]<a href="/cdn-cgi/l/email-protection" data-cfemail="0835487c6f7c">[email protected]</a>_ver -- sql server 2017
and p2.[version]<a href="/cdn-cgi/l/email-protection" data-cfemail="3c017c4f4e5f">[email protected]</a>_ver -- sql server 2012
)
select [Object]=p1.[schema]+'.'+p1.[Object]
, p1.[param], p1.param_type, p1.max_length,p1.is_output, p1.[version]
from dbo.AllParam p1
inner join c
on c.id=p1.id
left join dbo.AllParam p2
on p2.[schema]=p1.[schema] and p2.object = p1.object
and p2.[param]=p1.[param]
and p2.[version]<a href="/cdn-cgi/l/email-protection" data-cfemail="1d205d6e6f7e">[email protected]</a>_ver
where p2.[param] is null;
go

We can see the following changes:

Step 4 Test new objects

In the new stored procedure / functions, I am pretty interested in the 3 new extended stored procedures:

xp_copy_file
xp_copy_files
xp_delete_files

Before SQL Server 2019, for all copy and delete operations, we needed to rely on xp_cmdshell with embedded copy and delete commands, but with these three new SPs, we can do our work in a more T-SQL native way.

The basic syntax is:

-- NO wildcard allowed
exec master.sys.xp_copy_file 'c:\test\a.txt' -- source
, 'c:\temp\a1.txt' destination
-- Wildcard allowed
exec master.sys.xp_copy_files 'c:\test\a*.txt' -- source
, 'c:\temp\' destination
-- Wildcard allowed
exec master.sys.xp_delete_files 'c:\test\a*.txt' -- source

There is another interesting new table valued function, sys.dm_db_page_info, it has the following syntax

Sys.dm_db_page_info(<db_id>, <file_id>, <page num>, 'option') where [option] can be of ‘limited’ or ‘detailed’, and <page num> starts from 0.

This view will surely be an important tool for internal storage troubleshooting. A quick example is like the following:

select * from sys.dm_db_page_info(db_id('master'), 1, 1, 'detailed')

I get the following:

In theory, we can loop through each page of a database file via this view and then do analysis. For example, after an index rebuild, check how IAM s