Please support our sponsors
PASS Summit Annual international conference 6 to 9 November 2018, in Seattle (WA) 2 Days of pre-cons More than 200 sessions over 3 days Over 5,000 data platform professionals Evening networking activities I take liberties with these slides Ask about this
Some things to consider Please silence your mobile devices Please ask me lots of questions Please don’t snore 📖📖 Please buy my book
What’s new in SQL Server 2017? Randolph West Data Platform MVP BornSQL.ca r@ndolph.ca What’s new in SQL Server 2017?
Ok, but what’s new in SQL Server 2016?
Query Store The flight data recorder for your database
Security features Always Encrypted ✅ Dynamic Data Masking 🚫 Row Level Security 🤫
Temporal tables Keep track of data changes at table level
JSON support Like XML support, but better
In-memory improvements Making it actually useful: 2 TB max size ALTER TABLE / ALTER PROCEDURE / JOIN / UNION / DISTINCT / unique and foreign key constraints Supports blobs Supports TDE and other security Better collation options (not just BIN2) Parallel plans
Cool, so what’s new in SQL Server 2017?
Multi-platform support Microsoft Windows Ubuntu Linux SUSE Enterprise Linux Red Hat Linux Docker containers: Windows macOS Linux
Hey, I just met you, so demo maybe
Resumable online index rebuilds Shorter maintenance windows (max duration) Survives restarts Supports MAXDOP Supports log truncation Think of it as the ease of index defragmentation with the benefits of index rebuilds
Clusterless availability groups Readable secondaries? Try read-scale AGs. Migrate from Windows to Linux with no downtime.
Graph database Edges and nodes, oh my Many-to-many connections Examples include social media
Machine learning R and Python are first-class ML languages
Automatic database tuning In which the Query Store is used to force a plan
A very basic introduction* to joins - nested loop (small plus large data set) - merge (must be sorted, should be similar size) - hash (unsorted, for larger sets, lots of memory) * This is the one-minute version of a detailed topic and experts in the room will cringe at my oversimplification
Adaptive query processing batch mode adaptive joins memory grant feedback interleaved multi-line table-valued functions Requires a Columnstore index, but you can make an empty one and join to it, to get all the benefits
It’s already in Azure SQL Database