Azure Storage Tables

Understanding the Azure Storage Table Data Model

Azure Table storage is a NoSQL key-attribute store that allows you to store large amounts of structured, non-relational data. It's designed for scenarios where you need flexible data schemas and high scalability. This document explains the core concepts of the Azure Table storage data model.

Core Concepts

The fundamental entities in Azure Table storage are:

• Tables: A collection of entities.
• Entities: A set of properties, analogous to a row in a database.
• Properties: A name-value pair within an entity.

Tables

A table is a container for entities. All entities within a table share a similar structure (schema), but this structure can vary from entity to entity. You can have multiple tables within a storage account.

Entities

An entity is the basic unit of data in Azure Table storage. Each entity is a collection of properties. Every entity MUST have the following two properties:

• PartitionKey: A string value that dictates the logical grouping of entities. Entities with the same PartitionKey are stored together, which helps with query performance and scalability.
• RowKey: A string value that uniquely identifies an entity within a partition.

Together, PartitionKey and RowKey form the entity's unique identifier, known as the entity key.

Properties

In addition to PartitionKey and RowKey, an entity can contain up to 252 additional properties. Properties are name-value pairs.

• Property Names: Must be strings up to 255 characters in length. They are case-insensitive for querying but case-sensitive for storage.
• Property Values: Can be of various data types, including:
  • String
  • Int32
  • Int64
  • Double
  • DateTime
  • Boolean
  • Guid
  • Binary (Byte array)
  • DateTimeOffset
  • GeographyPoint (for Geo-JSON)

All properties are stored as Strings or DateTimes internally, and are converted to the appropriate .NET type when retrieved. Azure Table storage does not enforce a schema for these 252 properties. An entity can have a completely different set of properties from another entity in the same table.

Entity Keys (PartitionKey and RowKey)

The PartitionKey and RowKey are critical for efficient data access and scalability.

• Partitioning: Data is distributed across storage nodes based on the PartitionKey. This allows for massive scalability by distributing requests across multiple partitions.
• Querying: Queries that filter on PartitionKey are highly efficient as they target a specific partition. Queries across multiple partitions or without a PartitionKey filter are less performant.
• Uniqueness: The combination of PartitionKey and RowKey must be unique within a table.

Example Entity Structure

Operations on Entities

Azure Table storage supports the following operations on entities:

• Insert: Add a new entity.
• Update: Modify an existing entity. This can be done using an Update (replace) or Merge (partial update) operation.
• Delete: Remove an entity.
• Query: Retrieve entities based on filters.

Batch Operations

You can perform batch operations to insert, update, or delete multiple entities in a single request. However, batch operations are limited to entities within the same PartitionKey.

Data Types and Limits

Azure Table storage offers a flexible data model. However, be aware of the following limits:

• Maximum entity size: 1 MB (including property names and values).
• Maximum number of entities per table: No strict limit, dependent on storage account capacity.
• Maximum number of properties per entity: 255 (including PartitionKey and RowKey).
• Property name length: Up to 255 characters.
• Property value length: Varies by data type, with a maximum of 64KB for strings.