https://blogs.msdn.microsoft.com/sqlreleaseservices/sql-server-2016-service-pack-1-sp1-released/

https://docs.microsoft.com/en-us/sql/sql-server/editions-and-components-of-sql-server-2017?view=sql-server-2017

Data virtualization and SQL Server 2019 Big Data Clusters

New feature or update                             Details

Scalable big data solution                   Deploy scalable clusters of SQL Server, Spark, and HDFS containers running on Kubernetes.

                             Read, write, and process big data from Transact-SQL or Spark.

                                                            Easily combine and analyze high-value relational data with high-volume big data.

                                                            Query external data sources.

                                                            Store big data in HDFS managed by SQL Server.

                                                            Query data from multiple external data sources through the cluster.

                                                            Use the data for AI, machine learning, and other analysis tasks.

                                                            Deploy and run applications in Big Data Clusters.

                                                            The SQL Server master instance provides high availability and disaster recovery for all databases by using Always  

                                                             On availability group technology.

Data virtualization with PolyBase        Query data from external SQL Server, Oracle, Teradata, MongoDB, and ODBC data sources with external tables, 

                                                              now with UTF-8 encoding support. For more information, see What is PolyBase?.

Intelligent Query Processing

New feature or update                             Details

Row mode memory grant feedback             Expands on the batch mode memory grant feedback feature by adjusting memory grant sizes for both batch and row       mode operators. This adjustment can automatically correct excessive grants, which result in wasted memory and   

                                                                       reduced concurrency. It can also correct insufficient memory grants that cause expensive spills to disk. See Row  

                                                                 mode  memory grant feedback. 

Batch mode on rowstore

Enables batch mode execution without requiring columnstore indexes. Batch mode execution uses CPU more efficiently during analytical workloads but, until SQL Server 2019 (15.x), it was used only when a query included operations with columnstore indexes. However, some applications might use features that aren't supported with columnstore indexes and, therefore, can't leverage batch mode. Starting with SQL Server 2019 (15.x), batch mode is enabled on eligible analytical workloads whose queries include operations with any type of index (rowstore or columnstore). See Batch mode on rowstore.

Scalar UDF Inlining

Automatically transforms scalar UDFs into relational expressions and embeds them in the calling SQL query. This transformation improves the performance of workloads that take advantage of scalar UDFs. See Scalar UDF Inlining.

Table variable deferred compilation

Improves plan quality and overall performance for queries that reference table variables. During optimization and initial compilation, this feature propagates cardinality estimates that are based on actual table variable row counts. This accurate row count information optimizes downstream plan operations. See Table variable deferred compilation.

Approximate query processing with APPROX_COUNT_DISTINCT

For scenarios when absolute precision isn't important but responsiveness is critical, APPROX_COUNT_DISTINCT aggregates across large datasets while using fewer resources than COUNT(DISTINCT()) for superior concurrency. See Approximate query processing.