Query on database
The latest version of Microsoft SQL Server Database Engine Tuning Advisor (DTA) supports two new features: (a) Ability to recommend columnstore indexes in addition to rowstore indexes, and (b) Ability to automatically select a workload to tune from the Query Store. These new features are available when tuning for SQL Server 2016 Database Engine (or later) versions.
Recommendations for columnstore indexes
Data warehousing and analytic workloads often need to scan large amounts of data, and can greatly benefit from columnstore indexes. In contrast, rowstore (B+-tree) indexes are most effective for queries that access relatively small amounts of data searching for a particular range of values. Since rowstore indexes can deliver rows in sorted order, they can also reduce the cost of sorting in query execution plans. Therefore, the choice of which rowstore and columnstore indexes to build for your database is dependent on your application’s workload.
The latest version of DTA can analyze the workload and recommend a suitable combination of rowstore and columnstore indexes to build on the tables referenced by the workload. This article highlights the performance improvements achieved on real customer workloads by using DTA to recommend a combination of rowstore and columnstore indexes.
Tune Database using Workload from SQL Server Query Store
The Query Store feature in SQL Server automatically captures a history of queries, plans, and runtime statistics, and persists this information along with a database. It stores query execution statistics summarized over time intervals so you can see database usage patterns and understand when query plan changes happened on the server. DTA now supports a new option to analyze the Query Store to automatically select an appropriate workload for tuning. For many DTA users, this can take away the burden of having to collect a suitable workload file using SQL Server Profiler. This feature is only available if the database has Query Store turned on.