Merge pull request #32328 from MicrosoftDocs/main

OOB Publish for PASS Summit

Author: MikeRayMSFT

azure-sql/database/ai-artificial-intelligence-intelligent-applications.md

@@ -4,7 +4,7 @@ description: "Use AI options such as OpenAI and vectors to build intelligent app
 author: damauri
 ms.author: damauri
 ms.reviewer: damauri, josephsack, randolphwest, mathoma
-ms.date: 08/01/2024
+ms.date: 10/15/2024
 ms.service: azure-sql-database
 ms.topic: conceptual
 ms.collection: ce-skilling-ai-copilot
@@ -75,7 +75,7 @@ In Azure OpenAI, input text provided to the API is turned into tokens (tokenized
 
 ### Vectors
 
-Vectors are ordered arrays of numbers (typically floats) that can represent information about some data. For example, an image can be represented as a vector of pixel values, or a string of text can be represented as a vector or ASCII values. The process to turn data into a vector is called *vectorization*.
+Vectors are ordered arrays of numbers (typically floats) that can represent information about some data. For example, an image can be represented as a vector of pixel values, or a string of text can be represented as a vector or ASCII values. The process to turn data into a vector is called *vectorization*. For more information, see [Vectors](#vectors-1).
 
 ### Embeddings
 
@@ -89,7 +89,7 @@ Vector search refers to the process of finding all vectors in a dataset that are
 
 Consider a scenario where you run a query over millions of document to find the most similar documents in your data. You can create embeddings for your data and query documents using Azure OpenAI. Then, you can perform a vector search to find the most similar documents from your dataset. However, performing a vector search across a few examples is trivial. Performing this same search across thousands, or millions, of data points becomes challenging. There are also trade-offs between exhaustive search and approximate nearest neighbor (ANN) search methods including latency, throughput, accuracy, and cost, all of which depends on the requirements of your application.
 
-Since Azure SQL Database embeddings can be efficiently stored and queried using to columnstore index support, allowing exact nearest neighbor search with great performance, you don't have to decide between accuracy and speed: you can have both. Storing vector embeddings alongside the data in an integrated solution minimizes the need to manage data synchronization and accelerates your time-to-market for AI application development.
+Vectors in Azure SQL Database can be efficiently stored and queried, as described in the next sections, allowing exact nearest neighbor search with great performance. You don't have to decide between accuracy and speed: you can have both. Storing vector embeddings alongside the data in an integrated solution minimizes the need to manage data synchronization and accelerates your time-to-market for AI application development.
 
 ## Azure OpenAI
 
@@ -123,7 +123,33 @@ For additional examples on using SQL Database and OpenAI, see the following arti
 
 ## Vectors
 
-Although Azure SQL Database doesn't have a native **vector** type, a vector is nothing more than an ordered tuple, and relational databases are great at managing tuples. You can think of a tuple as the formal term for a row in a table.
+### Vector data type
+
+In November 2024, the new **vector** data type was introduced in Azure SQL Database.
+
+The dedicated **vector** type allows for efficient and optimized storing of vector data, and comes with a set of functions to help developers streamline vector and similarity search implementation. Calculating distance between two vectors can be done in one line of code using the new `VECTOR_DISTANCE` function. For more information on the [**vector** data type](/sql/t-sql/data-types/vector-data-type) and related functions, see [Overview of vectors in the SQL Database Engine](/sql/relational-databases/vectors/vectors-sql-server).
+
+For example:
+
+```sql
+CREATE TABLE [dbo].[wikipedia_articles_embeddings_titles_vector]
+(
+    [article_id] [int] NOT NULL,
+    [embedding] [vector](1536) NOT NULL,    
+)
+GO
+
+SELECT TOP(10) 
+    * 
+FROM 
+    [dbo].[wikipedia_articles_embeddings_titles_vector]
+ORDER BY
+    VECTOR_DISTANCE('cosine', @my_reference_vector, embedding)
+```
+
+### Vectors in older versions of SQL Server
+
+While older versions of SQL Server engine, up to and including SQL Server 2022, doesn't have a native **vector** type, a vector is nothing more than an ordered tuple, and relational databases are great at managing tuples. You can think of a tuple as the formal term for a row in a table.
 
 Azure SQL Database also supports columnstore indexes and [batch mode execution](/sql/relational-databases/query-processing-architecture-guide#batch-mode-execution). A vector-based approach is used for batch mode processing, which means that each column in a batch has its own memory location where it's stored as a vector. This allows for faster and more efficient processing of data in batches.
 
@@ -169,16 +195,18 @@ For an end-to-end sample to build a AI-enabled application using sessions abstra
 
 ### LangChain integration
 
-LangChain is a well-known framework for developing applications powered by language models.
+LangChain is a well-known framework for developing applications powered by language models. For examples that show how LangChain can be used to create a Chatbot on your own data, see:
 
-For an example that shows how LangChain can be used to create a Chatbot on your own data, see [Building your own DB Copilot for Azure SQL with Azure OpenAI GPT-4](https://devblogs.microsoft.com/azure-sql/building-your-own-db-copilot-for-azure-sql-with-azure-openai-gpt-4/).
+- [Build a chatbot on your own data in 1 hour with Azure SQL, Langchain and Chainlit](https://devblogs.microsoft.com/azure-sql/build-a-chatbot-on-your-own-data-in-1-hour-with-azure-sql-langchain-and-chainlit/): Build a chatbot using the RAG pattern on your own data using Langchain for orchestrating LLM calls and Chainlit for the UI.
+- [Building your own DB Copilot for Azure SQL with Azure OpenAI GPT-4](https://devblogs.microsoft.com/azure-sql/building-your-own-db-copilot-for-azure-sql-with-azure-openai-gpt-4/): Build a copilot-like experience to query your databases using natural language.
 
 ### Semantic Kernel integration
 
 [Semantic Kernel is an open-source SDK](/semantic-kernel/overview/) that lets you easily build agents that can call your existing code. As a highly extensible SDK, you can use Semantic Kernel with models from OpenAI, Azure OpenAI, Hugging Face, and more! By combining your existing C#, Python, and Java code with these models, you can build agents that answer questions and automate processes.
 
-- [Semantic Kernel & Kernel Memory - SQL Connector](https://github.com/kbeaugrand/SemanticKernel.Connectors.Memory.SqlServer) - Provides a connection to a SQL database for the Semantic Kernel for the memories.
+- [The ultimate chatbot?](https://devblogs.microsoft.com/azure-sql/the-ultimate-chatbot/): Build a chatbot on your own data using both NL2SQL and RAG patterns for the ultimate user experience. 
 - [OpenAI Embeddings Sample](https://github.com/marcominerva/OpenAIEmbeddingSample): An example that shows how to use Semantic Kernel and Kernel Memory to work with embeddings in a .NET application using SQL Server as Vector Database.
+- [Semantic Kernel & Kernel Memory - SQL Connector](https://github.com/kbeaugrand/SemanticKernel.Connectors.Memory.SqlServer) - Provides a connection to a SQL database for the Semantic Kernel for the memories.
 
 ## Microsoft Copilot skills in Azure SQL Database
 

azure-sql/database/doc-changes-updates-release-notes-whats-new.md

@@ -6,7 +6,7 @@ author: WilliamDAssafMSFT
 ms.author: wiassaf
 ms.reviewer: mathoma, randolphwest
 ms.service: azure-sql-database
-ms.date: 10/16/2024
+ms.date: 10/22/2024
 ms.subservice: service-overview
 ms.topic: whats-new
 ms.custom:
@@ -59,6 +59,7 @@ The following table lists the features of Azure SQL Database that are currently
 | [Query editor in the Azure portal](query-editor.md) | The query editor in the portal allows you to run queries against your Azure SQL Database directly from the [Azure portal](https://portal.azure.com). |
 | [SQL Analytics](/azure/azure-monitor/insights/azure-sql) | Azure SQL Analytics is an advanced cloud monitoring solution for monitoring performance of all of your Azure SQL databases at scale and across multiple subscriptions in a single view. Azure SQL Analytics collects and visualizes key performance metrics with built-in intelligence for performance troubleshooting. |
 | [UNISTR (Transact-SQL)](/sql/t-sql/functions/unistr-transact-sql) | Azure SQL Database now supports the `UNISTR` T-SQL syntax for Unicode string literals. For more information, see [UNISTR (Transact-SQL)](/sql/t-sql/functions/unistr-transact-sql).|
+| [Vector data type (preview)](/sql/relational-databases/vectors/vectors-sql-server) | Working with vector data is now easier in Azure SQL Database with the introduction of a new [vector data type](/sql/t-sql/data-types/vector-data-type) and [functions](/sql/t-sql/functions/vector-functions-transact-sql). For more information, see [Intelligent applications with Azure SQL Database](ai-artificial-intelligence-intelligent-applications.md#vectors).|
 | [\|\|](/sql/t-sql/language-elements/string-concatenation-pipes-transact-sql) and [\|\|=](/sql/t-sql/language-elements/compound-assignment-pipes-transact-sql) syntax support | Azure SQL Database now supports [\|\| (String concatenation)](/sql/t-sql/language-elements/string-concatenation-pipes-transact-sql) and [\|\|= (Compound assignment)](/sql/t-sql/language-elements/compound-assignment-pipes-transact-sql) Transact-SQL syntax.|
 
 ## General availability (GA)
@@ -76,7 +77,7 @@ The following table lists features of Azure SQL Database that have been made gen
 | [Automatic backups on secondary replicas](automated-backups-overview.md#automatic-backups-on-secondary-replicas) | August 2024 | Mitigate the performance impact on your workload by taking automated backups from the non-readable secondary replica in the Business Critical service tier. |
 | [Database compatibility level 160 is now default](/sql/t-sql/statements/alter-database-transact-sql-compatibility-level?view=azuresqldb-current&preserve-view=true) | June 2024 | Database compatibility level 160 is now the default for new databases created in Azure SQL Database. For more information on this announcement, see [General availability: Database compatibility level 160 in Azure SQL Database](https://techcommunity.microsoft.com/t5/azure-sql-blog/general-availability-database-compatibility-level-160-in-azure/ba-p/4172039). |
 | [Hyperscale named replica zone redundant support](service-tier-hyperscale-replicas.md) | June 2024 | [Zone redundancy support for Hyperscale named replicas](https://aka.ms/ZRSupportForNRPreview) is now generally available. |
-| [License-free standby replica](standby-replica-how-to-configure.md) | May 2024 | Save on licensing costs by configuring your secondary database replica for disaster recovery standby. | 
+| [License-free standby replica](standby-replica-how-to-configure.md) | May 2024 | Save on licensing costs by configuring your secondary database replica for disaster recovery standby. |
 | [Elastic jobs](elastic-jobs-overview.md) | April 2024 | [Elastic jobs, now generally available](https://techcommunity.microsoft.com/t5/azure-sql-blog/general-availability-elastic-jobs-in-azure-sql-database/ba-p/4087140), are the SQL Server Agent replacement for Azure SQL Database. Elastic jobs  support Microsoft Entra ID authentication, private endpoints, management via REST APIs, Azure Alerts, and more new features since public preview began. |
 | [Maintenance window advance notifications](advance-notifications.md) | March 2024 | Advance notifications are now generally available for databases configured to use a nondefault [maintenance window](maintenance-window.md). |
 | [Azure SQL triggers for Azure Functions](/azure/azure-functions/functions-bindings-azure-sql-trigger) | March 2024 | Azure Functions supports function triggers for Azure SQL Database. |
@@ -93,6 +94,7 @@ Learn about significant changes to the Azure SQL Database documentation. For pre
 
 | Changes | Details |
 | --- | --- |
+| **Vector data type (preview)** | Working with vector data is now easier in Azure SQL Database with the introduction of a new [vector data type](/sql/t-sql/data-types/vector-data-type) and [functions](/sql/t-sql/functions/vector-functions-transact-sql). For more information, see [Intelligent applications with Azure SQL Database](ai-artificial-intelligence-intelligent-applications.md#vectors).|
 | **Hyperscale single database increased maximum size** | The maximum single database size in Azure SQL Database Hyperscale has been increased from 100 TB to 128 TB. For more information, see [Blog: November 2024 Hyperscale enhancements](https://aka.ms/AAslnql).|
 | **Hyperscale increased log generation rate (preview)**| The transaction log generation rate in Azure SQL Database Hyperscale single databases is set to increase from 100 MB/s to 150 MB/s. The increased log generation rate is available as an opt-in preview feature. For more information and to opt-in to 150 MB/s, see [Blog: November 2024 Hyperscale enhancements](https://aka.ms/AAslnql).|
 | **Hyperscale continuous priming (preview)**| [Continuous priming](service-tier-hyperscale.md#buffer-pool-resilient-buffer-pool-extension-and-continuous-priming) is an innovative new feature is designed to optimize Hyperscale performance during failovers by priming secondary compute replicas. Continuous priming is currently in a gated preview. For more information and to opt-in to continuous priming, see [Blog: November 2024 Hyperscale enhancements](https://aka.ms/AAslnql).|
@@ -101,7 +103,7 @@ Learn about significant changes to the Azure SQL Database documentation. For pre
 
 | Changes | Details |
 | --- | --- |
-| **Lower auto-pause delay for serverless** | Reduce costs by lowering the [auto-pause delay for serverless compute in Azure SQL Database](https://aka.ms/AAs7lpz). |
+| **Lower auto-pause delay for serverless** | Reduce costs by lowering the [auto-pause delay for serverless compute in Azure SQL Database](https://aka.ms/AAs7lpz). For more information, see [serverless compute tier](serverless-tier-overview.md).|
 
 ### September 2024
 

docs/relational-databases/vectors/vectors-sql-server.md

@@ -0,0 +1,81 @@
+---
+title: Work with vectors in the SQL Database Engine
+description: How to create, manage, and search vectors in the SQL Database Engine.
+author: WilliamDAssafMSFT
+ms.author: wiassaf
+ms.reviewer: damauri, pookam, jovanpop, randolphwest
+ms.date: 10/22/2024
+ms.service: sql
+ms.topic: conceptual
+ms.custom:
+  - intro-quickstart
+helpviewer_keywords:
+  - "Vectors"
+  - "Vectors, built-in support"
+monikerRange: "=azuresqldb-current"
+---
+# Overview of vectors in the SQL Database Engine
+
+[!INCLUDE [Azure SQL Database](../../includes/applies-to-version/asdb.md)]
+
+Vectors are ordered arrays of numbers (typically floats) that can represent information about some data. For example, an image can be represented as a vector of pixel values, or a string of text can be represented as a vector or ASCII values. The process to turn data into a vector is called vectorization.
+
+## Embeddings
+
+Embeddings are vectors that represent important features of data. Embeddings are often learned by using a deep learning model, and machine learning and AI models utilize them as features. Embeddings can also capture semantic similarity between similar concepts. For example, in generating an embedding for the words `person` and `human`, we would expect their embeddings (vector representation) to be similar in value since the words are also semantically similar.
+
+Azure OpenAI features models to create embeddings from text data. The service breaks text out into tokens and generates embeddings using models pretrained by OpenAI. To learn more, see [Creating embeddings with Azure OpenAI](/azure/ai-services/openai/concepts/understand-embeddings).
+
+Once embeddings are generated, they can be stored into a SQL Server database. This allows you to store the embeddings alongside the data they represent, and to perform vector search queries to find similar data points.
+
+## Vector search
+
+Vector search refers to the process of finding all vectors in a dataset that are similar to a specific query vector. Therefore, a query vector for the word `human` searches the entire dataset for similar vectors, and thus similar words: in this example it should find the word `person` as a close match. This closeness, or distance, is measured using a distance metric such as cosine distance. The closer vectors are, the more similar they are.
+
+SQL Server provides built-in support for vectors via the **vector** data type. Vectors are stored in an optimized binary format but exposed as JSON arrays for convenience. Each element of the vector is stored using single-precision (4 bytes) floating-point value. Along with the data type there are dedicated functions to operate on vectors. For example, it's possible to find the distance between two vectors using the `VECTOR_DISTANCE` function. The function returns a scalar value with the distance between two vectors based on the distance metric you specify.
+
+Since vectors are typically managed as arrays of floats, creating a vector can be done simply casting a JSON array to a **vector** data type. For example, the following code creates a vector from a JSON array:
+
+```sql
+SELECT CAST('[1.0, -0.2, 30]' AS VECTOR(3)) AS vector;
+```
+
+or, using implicit casting
+
+```sql
+DECLARE @v VECTOR(3) = '[1.0, -0.2, 30]';
+SELECT @v;
+```
+
+Same goes for converting a vector into a JSON array:
+
+```sql
+DECLARE @v VECTOR(3) = '[1.0, -0.2, 30]';
+SELECT CAST(@v AS NVARCHAR(MAX)) AS vector;
+```
+
+## Limitations
+
+In the current preview casting to and from JSON data type is not supported yet. The workaround is to first convert from/to **NVARCHAR(MAX)** and then to/from JSON. For example, to convert a vector to a JSON type:
+
+```sql
+DECLARE @v VECTOR(3) = '[1.0, -0.2, 30]';
+SELECT CAST(CAST(@v AS NVARCHAR(MAX)) AS JSON) AS j;
+```
+
+and to convert from a JSON type to vector:
+
+```sql
+DECLARE @j JSON = JSON_ARRAY(1.0, -0.2, 30)
+SELECT CAST(CAST(@j AS NVARCHAR(MAX)) AS VECTOR(3)) AS v;
+```
+
+More details on how to use vectors in SQL Server can be found in the following articles:
+
+- [Vector Data Types](../../t-sql/data-types/vector-data-type.md)
+- [Vector Functions](../../t-sql/functions/vector-functions-transact-sql.md)
+- [Azure SQL DB Vector Search Samples](https://github.com/Azure-Samples/azure-sql-db-vector-search)
+
+## Related content
+
+- [Intelligent applications with Azure SQL Database](/azure/azure-sql/database/ai-artificial-intelligence-intelligent-applications)

docs/t-sql/data-types/data-types-transact-sql.md

@@ -110,16 +110,17 @@ In [!INCLUDE [ssNoVersion](../../includes/ssnoversion-md.md)], based on their st
 
 ### Other data types
 
-- [cursor](cursor-transact-sql.md)
+- [cursor](cursor-transact-sql.md)
 - [geography](../spatial-geography/spatial-types-geography.md) <sup>1</sup>
-- [geometry](../spatial-geometry/spatial-types-geometry-transact-sql.md) <sup>1</sup>
+- [geometry](../spatial-geometry/spatial-types-geometry-transact-sql.md) <sup>1</sup>
 - [hierarchyid](hierarchyid-data-type-method-reference.md)
 - [json](json-data-type.md)
-- [rowversion](rowversion-transact-sql.md)
-- [sql_variant](sql-variant-transact-sql.md)
-- [table](table-transact-sql.md)
-- [uniqueidentifier](uniqueidentifier-transact-sql.md)
-- [xml](../xml/xml-transact-sql.md)
+- [vector](vector-data-type.md)
+- [rowversion](rowversion-transact-sql.md)
+- [sql_variant](sql-variant-transact-sql.md)
+- [table](table-transact-sql.md)
+- [uniqueidentifier](uniqueidentifier-transact-sql.md)
+- [xml](../xml/xml-transact-sql.md)
 
 <sup>1</sup> The **geography** and **geometry** data types are *spatial types*.