Merge branch 'master' of https://github.com/MicrosoftDocs/sql-docs-pr into release-qds-hints

Author: v-alje

.openpublishing.redirection.json

@@ -42945,6 +42945,21 @@
       "redirect_url": "/sql/machine-learning/tutorials/python-ski-rental-linear-regression-deploy-model",
       "redirect_document_id": false
     },
+    {
+      "source_path": "docs/machine-learning/r/save-and-load-r-objects-from-sql-server-using-odbc.md",
+      "redirect_url": "/sql/machine-learning/data-exploration/save-and-load-r-objects-from-sql-server-using-odbc",
+      "redirect_document_id": true
+    },
+    {
+      "source_path": "docs/machine-learning/r/creating-multiple-models-using-rxexecby.md",
+      "redirect_url": "/sql/machine-learning/data-exploration/creating-multiple-models-using-rxexecby",
+      "redirect_document_id": true
+    },
+    {
+      "source_path": "docs/machine-learning/r/data-exploration-and-predictive-modeling-with-r.md",
+      "redirect_url": "/sql/machine-learning/tutorials/r-tutorials",
+      "redirect_document_id": false
+    },
     {
       "source_path": "docs/linux/sql-server-linux-develop-use-ssms.md",
       "redirect_url": "/sql/linux/sql-server-linux-manage-ssms",

docs/connect/ado-net/appcontext-switches.md

@@ -1,7 +1,7 @@
 ---
 title: AppContext switches in SqlClient
 description: Learn about the AppContext switches available in SqlClient and how to use them to modify some default behaviors.
-ms.date: 03/24/2021
+ms.date: 05/28/2021
 dev_langs: 
   - "csharp"
 ms.prod: sql
@@ -32,6 +32,8 @@ AppContext.SetSwitch("Switch.Microsoft.Data.SqlClient.TruncateScaledDecimal", tr
 
 [!INCLUDE [appliesto-xxxx-netcore-netst-md](../../includes/appliesto-xxxx-netcore-netst-md.md)]
 
+(Available starting with version 2.0)
+
 On Windows, SqlClient uses a native implementation of the SNI network interface by default. To enable the use of a managed SNI implementation, you can set the AppContext switch **"Switch.Microsoft.Data.SqlClient.UseManagedNetworkingOnWindows"** to `true` at application startup:
 
 ```csharp
@@ -92,6 +94,8 @@ AppContext.SetSwitch("Switch.Microsoft.Data.SqlClient.MakeReadAsyncBlocking", fa
 
 [!INCLUDE [appliesto-netfx-netcore-netst-md](../../includes/appliesto-netfx-netcore-netst-md.md)]
 
+(Available starting with version 3.0)
+
 By default, configurable retry logic is disabled. To enable this feature, set the AppContext switch **Switch.Microsoft.Data.SqlClient.EnableRetryLogic** to `true` at application startup. This switch is required, even if a retry provider is assigned to a connection or command.
 
 ```csharp

docs/includes/synapse-analytics-od-unsupported-syntax.md

@@ -1,2 +1 @@
-
 This syntax is not supported by serverless SQL pool in Azure Synapse Analytics.

…eating-multiple-models-using-rxexecby.md …eating-multiple-models-using-rxexecby.mddocs/machine-learning/r/creating-multiple-models-using-rxexecby.md renamed to docs/machine-learning/data-exploration/creating-multiple-models-using-rxexecby.md docs/machine-learning/r/creating-multiple-models-using-rxexecby.md renamed to docs/machine-learning/data-exploration/creating-multiple-models-using-rxexecby.md

@@ -3,26 +3,27 @@ title: Create multiple models with rxExecBy
 description: Use the rxExecBy function from RevoScaleR library to build multiple mini models over machine data stored in SQL Server.
 ms.prod: sql
 ms.technology: machine-learning-services
-
-ms.date: 04/15/2018  
+ms.date: 05/28/2021
 ms.topic: how-to
 author: dphansen
 ms.author: davidph
-ms.custom: seo-lt-2019
+ms.custom: contperf-fy21q4
 monikerRange: ">=sql-server-2016||>=sql-server-linux-ver15"
 ---
 # Creating multiple models using rxExecBy
 [!INCLUDE [SQL Server 2016 and later](../../includes/applies-to-version/sqlserver2016.md)]
 
-The **rxExecBy** function in RevoScaleR supports parallel processing of multiple related models. Rather than train one large model based on data from multiple similar entities, a data scientist can quickly create many related models, each using data specific to a single entity. 
+Learn how to use the **rxExecBy** function in RevoScaleR to parallel process multiple related models with [SQL Server Machine Learning Services](../sql-server-machine-learning-services.md). Rather than train one large model based on data from multiple similar entities, you can quickly create many related models, each using data specific to a single entity.
+
+## What rxExecBy can do
 
 For example, suppose you are monitoring device failures, capturing data for many different types of equipment. By using rxExecBy, you can provide a single large dataset as input, specify a column on which to stratify the dataset, such as device type, and then create multiple models for individual devices.
 
 This use case has been termed ["pleasingly parallel"](https://en.wikipedia.org/wiki/Embarrassingly_parallel) because it breaks a large complicated problem into component parts for concurrent processing.
 
 Typical applications of this approach include forecasting for individual household smart meters, creating revenue projections for separate product lines, or creating models for loan approvals that are tailored to individual bank branches.
 
-## How rxExec Works
+## How rxExecBy works
 
 The rxExecBy function in RevoScaleR is designed for high-volume parallel processing over a large number of small data sets.
 
@@ -32,7 +33,7 @@ The rxExecBy function in RevoScaleR is designed for high-volume parallel process
 4. When the function executes, the data queries are processed in parallel if your environment supports it. Moreover, the modeling or transformation tasks are distributed among individual cores and executed in parallel. Supported compute context for thee operations include RxSpark and RxInSQLServer.
 5. Multiple results are returned.
 
-## rxExecBy Syntax and Examples
+## rxExecBy syntax and examples
 
 **rxExecBy** takes four inputs, one of the inputs being a dataset or data source object that can be partitioned on a specified **key** column. The function returns an output for each partition. The form of the output depends on the function that is passed as an argument. For example, if you pass a modeling function such as rxLinMod, you could return a separate trained model for each partition of the dataset.
 
@@ -89,4 +90,6 @@ rxSetComputeContext(sqlServerCC)
 sqlServerCCResults <- rxExecBy(inData = sqlServerDataDS, keys = c("DayOfWeek"), func = .Count)
 ```
 
+## Next steps
 
++ [What is SQL Server Machine Learning Services?](../sql-server-machine-learning-services.md)

…-r-objects-from-sql-server-using-odbc.md …-r-objects-from-sql-server-using-odbc.mddocs/machine-learning/r/save-and-load-r-objects-from-sql-server-using-odbc.md renamed to docs/machine-learning/data-exploration/save-and-load-r-objects-from-sql-server-using-odbc.md docs/machine-learning/r/save-and-load-r-objects-from-sql-server-using-odbc.md renamed to docs/machine-learning/data-exploration/save-and-load-r-objects-from-sql-server-using-odbc.md

@@ -4,23 +4,21 @@ description: The RevoScaleR package includes serialization and deserialization f
 ms.prod: sql
 ms.technology: machine-learning-services
 
-ms.date: 04/15/2018  
+ms.date: 04/27/2021
 ms.topic: how-to
 author: dphansen
 ms.author: davidph
-ms.custom: seo-lt-2019
+ms.custom: contperf-fy21q4
 monikerRange: ">=sql-server-2016||>=sql-server-linux-ver15"
 ---
-# Save and load R objects from SQL Server using ODBC
+# Use ODBC to save and load R objects in SQL Server Machine Learning Services
 [!INCLUDE [SQL Server 2016 and later](../../includes/applies-to-version/sqlserver2016.md)]
 
-SQL Server R Services can store serialized R objects in a table and then load the object from the table as needed, without you having to re-run the R code or retrain the model. This ability to save R objects in a database is critical for scenarios such as training and saving a model, and then using it later for scoring or analysis.
+Learn how to use the **RevoScaleR** package to store serialized R objects in a table and then load the object from the table as needed with [SQL Server Machine Learning Services](../sql-server-machine-learning-services.md). This can be used when training and saving a model, and then use it later for scoring or analysis.
 
-To improve performance of this critical step, the **RevoScaleR** package now includes new serialization and deserialization functions that greatly improve performance, and store the object more compactly. This article describes these functions and how to use them.
+## RevoScaleR package
 
-## Overview
-
-The **RevoScaleR** package now includes new functions that make it easier to save R objects to SQL Server and then read the objects from the SQL Server table. In general, each function call uses a simple key value store, in which the key is the name of the object, and the value associated with the key is the varbinary R object to be moved in or out of a table.
+The **RevoScaleR** package includes serialization and deserialization functions that can R objects compactly to SQL Server and then read the objects from the table. In general, each function call uses a simple key value store, in which the key is the name of the object, and the value associated with the key is the varbinary R object to be moved in or out of a table.
 
 To save R objects to SQL Server directly from an R environment, you must:
 
@@ -40,7 +38,7 @@ By default, any object that you call from R to move to SQL Server is serialized
 
 - `rxListKeys` lists as key-value pairs all the available objects. This helps you determine the names and versions of the R objects.
 
-For detailed help on the syntax of each function, use R help. Details are also available in the [ScaleR reference](/r-server/r-reference/revoscaler/revoscaler).
+For detailed help on the syntax of each function, use R help. Details are also available in the [ScaleR reference](/machine-learning-server/r-reference/revoscaler/revoscaler).
 
 ## How to store R objects in SQL Server using ODBC
 
@@ -109,4 +107,8 @@ This procedure demonstrates how you can use the new functions to load a model fr
 
    ```R
     infertLogit2 <- rxReadObject(ds, "logit.model")
-   ```
+   ```
+
+## Next steps
+
++ [What is SQL Server Machine Learning Services?](../sql-server-machine-learning-services.md)

docs/machine-learning/r/data-exploration-and-predictive-modeling-with-r.md

@@ -247,11 +247,9 @@ items:
     - name: R
       items:
       - name: Load R objects using ODBC
-        href: ../machine-learning/r/save-and-load-r-objects-from-sql-server-using-odbc.md
-      - name: Predictive modeling with R
-        href: ../machine-learning/r/data-exploration-and-predictive-modeling-with-r.md
+        href: ../machine-learning/data-exploration/save-and-load-r-objects-from-sql-server-using-odbc.md
       - name: Create multiple models using rxExecBy
-        href: ../machine-learning/r/creating-multiple-models-using-rxexecby.md
+        href: ../machine-learning/data-exploration/creating-multiple-models-using-rxexecby.md
   - name: Data type conversions
     items:
     - name: Python to SQL