SQL Server 2022 Programmability & Performance

SQL
START!
2023
SQL Server 2022
Programmability & Performance
Gianluca Hotz
ghotz@ugiss.org | @glhotz

Who am I?
• Gianluca Hotz | @glhotz | ghotz@ugiss.org
• Independent Consultant
• 25+ years on SQL Server (from 4.21 back in 1996)
• Database modeling & development, sizing & administration, modernization
(upgrades & migrations), performance tuning, security
• Community
• 24 years Microsoft MVP SQL Server/Data Platform (from 1998)
• VMware Experts SQL Server
• Founder and president UGISS (ex «PASS Chapter»)

Agenda
• Programmability
• Performance

Programmability
SQL Server 2022 Programmability & Performance

T-SQL
• GREATEST
• LEAST
• STRING_SPLIT
• TRIM functions
• DATETRUNC
• IS [NOT] DISTINCT FROM
• WINDOW clause
• LEFT_SHIFT
• RIGHT_SHIFT
• BIT_COUNT
• GET_BIT
• SET_BIT
JSON
• ISJSON
• JSON_PATH_EXISTS
• JSON_OBJECT
• JSON_ARRAY
Time Series
• DATE_BUCKET
• GENERATE_SERIES
• FIRST_VALUE
• LAST_VALUE
Programmability enhancements

XML Compression
• Finally!!! 
• New CREATE/ALTER TABLE or CREATE/ALTER INDEX option
• XML_COMPRESSION = ON|OFF [ON PARTITIONS(…)]
• More info
• https://learn.microsoft.com/sql/t-sql/statements/create-index-transact-
sql#xml_compression

Multi-write replication
• Last Writer Wins (LWW) in Peer-to-peer transactional replication
• Introduced in SQL Server 2019 CU13
• Previously manual resolution in case of conflict and replication paused
• Enterprise Edition only
• More info
• https://learn.microsoft.com/sql/relational-
databases/replication/transactional/peer-to-peer-conflict-detection-in-peer-
to-peer-replication#automatically-handle-conflicts-with-last-write-wins
• https://techcommunity.microsoft.com/t5/sql-server-blog/replication-
enhancements-in-the-sql-server-2019-cu13-release/ba-p/2814727

Always Encrypted with Secure Enclaves
• Enhancements inside enclave
• Multiple threads and key caching
• More operations supported
• More info
• https://learn.microsoft.com/sql/relationa
l-databases/security/encryption/always-
encrypted-enclaves
Operation/
Support
Azure SQL
Database
SQL Server
2022
SQL Server
2019
Comparison
operators
Yes Yes Yes
BETWEEN Yes Yes Yes
IN Yes Yes Yes
LIKE Yes Yes Yes
DISTINCT Yes Yes Yes
Joins Yes Yes Only nested
loops
ORDER BY Yes Yes No
GROUP BY Yes Yes No

Append-Only
Ledger Table
Trusted
Storage
Database Ledger
Block N Block N-1 Block N-2
Updatable
Ledger Table
History Table
Ledger View
Ledger Tables
• Updatable allow insert/update/delete
• History of updated/deleted rows preserved in
history table and easy-to-query Ledger View
• Integrity of updatable/history tables maintained
through cryptographic links of the Database
Ledger
• System can periodically upload digital receipts to
a customer-configured trusted storage service
• Customer can use digital receipts to verify the
integrity of the data
• Append-Only allow only insert
• no need for a history table

HTTPS REST API
SQL Engine
wasb
abs
abfs
adls

Performance
SQL Server 2022 Programmability & Performance

Buffer Pool parallel scans
• BP operations use hash tables (BUF structures to find pages)
• Some operations still needs to scan all BUF structures
• Problem on systems with large amount of memory e.g. 1TB+
• SQL Server warning in ERRORLOG
• https://docs.microsoft.com/troubleshoot/sql/performance/buffer-pool-scan-runs-
slowly-large-memory-machines
• SQL Server 2022 introduces buffer pool parallel scan
• Enabled by default (available in Standard and Enterprise editions)
• Standard Edition limited to 2 threads on 64GB
• More info
• https://cloudblogs.microsoft.com/sqlserver/2022/07/07/improve-scalability-with-
buffer-pool-parallel-scan-in-sql-server-2022
• https://youtu.be/4GvU106Xiag

“Hands-free” tempdb
Pre
SQL
Server
2019
1 file = PFS, GAM,
SGAM contention
Add multiple files
Trace flags 1117
and 1118
SQL 2016 setup auto
adds multiple files
Trace flags not
required by SQL
Server 2016
SQL
Server
2022
SGAM and GAM
concurrency
Latch contention gone
SQL
Server
2019
PFS concurrency
Autogrow and
uniform default
for tempdb
Now system table
pages become
hotspot
Tempdb metadata
optimization ON
SGAM and GAM
contention remain

“Purvi’s list”
• Reduced buffer pool I/O promotions
• Tuned read-ahead to avoid single page promotions to 8 pages I/Os
• Enhanced spinlock algorithms
• No details, internal adjustments make spinlocks more efficient…
• Improved Virtual Log File (VLF) algorithms
• If growth > 1/8 current size, if < 64MB creates 1 VLF instead of 4
• https://learn.microsoft.comsql/relational-databases/sql-server-transaction-
log-architecture-and-management-guide#virtual-log-files-vlfs
• Instant file initialization (IFI) for transaction log file growth events!!!
• Only growth events up to 64MB!

Columnstore enhancements
• Ordered Clustered Columnstore Index!
• https://docs.microsoft.com/azure/synapse-analytics/sql-data-
warehouse/performance-tuning-ordered-cci
• Columnstore string enhancements
• Deep data (e.g. char, binary, guid) min/max maintained when rebuilding
• Fast string-equal operation
• LIKE pushdown (RG elimination only for prefix searches i.e. str% not %str)
• Segment elimination
• Extends to string, binary, guid data types and datetimeoffset w/ scale > 2
• Was only numeric, date, time data types and datetimeoffset w/ scale <= 2

Batch mode enhancements
• Processor Advanced Vector Extension (AVX) 512
• Some operations faster for Columnstore and Rowstore batch mode
• Recommended for the following processors
• Intel Ice Lake and later
• AMD EYPC Genoa and later
• Currently enabled by trace flag 15097
• https://learn.microsoft.com/sql/t-sql/database-console-commands/dbcc-
traceon-trace-flags-transact-sql#tf15097

Hybrid Buffer Pool
• SQL Server 2019
• Clean pages directly referenced on PMEM devices without copy
• Dirty pages still kept in DRAM
• SQL Server 2022
• Direct write, reduces number of memcpy operations
• Currently enabled by trace flag 809
• https://learn.microsoft.com/sql/t-sql/database-console-commands/dbcc-traceon-trace-
flags-transact-sql#tf809
• More info
• https://learn.microsoft.com/sql/database-engine/configure-windows/hybrid-
buffer-pool#hybrid-buffer-pool-with-direct-write

Query Store enhancements
• On by default (migrated databases retain original configuration)
• Query Store Hints (shape plans with no code changes)
• Query Store support for AG secondary replicas
• Now used also by Intelligent Query Processing

Query Store hints
• Process much simpler than Plan Guides
• Find query id in Query Store
• Use sys.sp_query_store_set_hints to apply hint
• Check for hint failures in sys.query_store_query_hints
• Hints survive plan cache eviction and restart!
• More info
• https://learn.microsoft.com/sql/relational-databases/performance/query-
store-hints

Query Store hints support
• Supported hints
• { HASH | ORDER } GROUP
• { CONCAT | HASH | MERGE } UNION
• { LOOP | MERGE | HASH } JOIN
• EXPAND VIEWS
• FAST number_rows
• FORCE ORDER
• IGNORE_NONCLUSTERED_COLUMNSTORE_INDEX
• KEEP PLAN
• KEEPFIXED PLAN
• MAX_GRANT_PERCENT = percent
• MIN_GRANT_PERCENT = percent
• MAXDOP number_of_processors
• NO_PERFORMANCE_SPOOL
• OPTIMIZE FOR UNKNOWN
• PARAMETERIZATION { SIMPLE | FORCED }
• RECOMPILE
• ROBUST PLAN
• USE HINT ( '<hint_name>' [ , ...n ] )
• Unsupported hints
• OPTIMIZE FOR(@var = val)
• MAXRECURSION
• USE PLAN
• consider QP original plan forcing
• DISABLE_DEFERRED_COMPILATION_TV
• DISABLE_TSQL_SCALAR_UDF_INLINING
• Table hints
• E.g. FORCESEEK, READUNCOMMITTED, INDEX
• Feature interoperability
• https://learn.microsoft.com/sql/relational-
databases/performance/query-store-hints#query-
store-hints-and-feature-interoperability

Query Store for AG secondary replicas
• Primary replica stores data
• For all replicas
• Column replica_group_id only in
• sys.query_store_replicas
• sys.query_store_runtime_stats
• sys.query_store_wait_stats
• Enabled by trace flag 12606
• More info
• https://learn.microsoft.com/sql/relationa
l-databases/performance/query-store-
for-secondary-replicas

Query Store and IQP
Query Store
Degree of Parallelism
(DOP) feedback
Parameter Sensitive Plan
(PSP) Optimization
Cardinality Estimation
(CE) feedback
Intelligent Query Processing (IQP)
Next Generation
Feedback store
Hints

Intelligent Query Processing Gen 3
Intelligent QP
Adaptive QP
Adaptive Joins
Batch Mode
Interleaved
Execution
Query
Processing
Feedback
Memory Grant
Feedback
Row Mode Batch Mode
Percentile
Grant
Feedback
CE Feedback DOP Feedback
Feedback
Persistence
Parameter
Sensitive Plan
Optimization
Optimized
Plan Forcing
Approximate
QP
Approximate
Count Distinct
Approximate
Percentile
Table Variable
Deferred
Compilation
Batch Mode
on Row Store
Scalar UDF
Inlining
SQL Server 2017+
SQL Server 2019+
SQL Server 2022+
Enterprise Edition

IPQ changes not gated by dbcompat…
SQL Server
2022
• Approximate Percentile
• PERCENTILE_CONT
• PERCENTILE_DISC
• Optimized Plan Forcing
• Persists compile steps
• Reduce compilation
overhead
dbcompat
140+
• Memory Grant
Percentiles
• Smooth oscillation
• Memory Grant Feedback
Persistence
• Survive cache eviction
and restart
dbcompat 160
• Parameter Sensitive Plan
(PSP) Optimization
• Cardinality Estimation
(CE) Feedback
• Degree of Parallelism
(DOP) feedback

Hardware offloaded backup compression
• Currently only Intel QuickAssist (QAT) supported
• Install drivers
• Diagnostic DMV sys.dm_server_accelerator_status
• Enable functionality
• sp_configure 'hardware offload enabled’
• Enable specific accelerator
• ALTER SERVER CONFIGURATION SET
HARDWARE_OFFLOAD = ON|OFF (ACCELERATOR = QAT[, MODE = SOFTWARE])
• Backup
• COMPRESSION ALGORITHM = MS_XPRESS|other_algorithm
• Usually, good performance with concurrent intensive workload
• More info
• https://docs.microsoft.com/sql/relational-databases/system-tables/backupset-
transact-sql

Learn more about SQL Server 2022
aka.ms/sqlserver2022
Download SQL Server 2022
aka.ms/getsqlserver2022
Get our decks
aka.ms/sqlserver2022decks
What’s new for SQL Server 2022
aka.ms/sqlserver2022docs
Try our demos
aka.ms/sqlserver2022demos
Watch Microsoft Mechanics
aka.ms/sqlserver2022mechanics
Don’t miss us on Data Exposed
aka.ms/dataexposed
Read the SQL Server 2022 Blog Series
aka.ms/sqlserver2022blogs
Take the Microsoft Learning Path
aka.ms/learnsqlserver2022
Read the book
aka.ms/sql2022book
Try out the workshop
aka.ms/sql2022workshop

SQL Server 2022 Programmability & Performance

More Related Content

Similar to SQL Server 2022 Programmability & Performance

More from Gianluca Hotz

Recently uploaded

SQL Server 2022 Programmability & Performance

Editor's Notes