Holds structured NoSQL records without a fixed schema, using keys and attributes; think device details, contact lists or user records. It now sits within the Azure Cosmos DB family, where Azure Cosmos DB for Table is the premium step up.
Read more: Microsoft Learn
In the Ultra Transcenders books
Each book explains Azure Table Storage in context, with comparison tables and the common traps.
Terms in this definition
- NoSQL
Document-oriented Cosmos DB API: data is JSON and queries use a dialect resembling SQL.
- Fixed schema
With the Replace update method, this Dataflow Gen2 option keeps the existing destination table and swaps only its contents, so any relationships and measures survive. Warehouse destinations offer nothing else; the Dynamic schema alternative drops and rebuilds the table.
- Chat message roles
Labels on chat messages: instructions go under system, the person's input under user, the model's previous answers under assistant, and results returned by a called tool under tool (or function).
- Cosmos DB
NoSQL database distributed globally, offering writes in multiple regions, automatic indexing and latency below 10 ms.
- WHERE
Limits a SELECT, UPDATE or DELETE to just the rows meeting a condition. Omit it, and the statement hits every row.
- Azure Cosmos DB for Table
Works like Azure Table Storage, with its key/value model and programming approach, but gains global distribution, higher throughput and indexing of every property. Microsoft suggests it when starting new table-based workloads.
- Premium
Hosting plan for Azure Functions that keeps instances pre-warmed to avoid cold starts and supports VNet integration. Executions time out after 30 minutes by default, which host.json can extend.
Related terms
- Cosmos DB for Table
API in Cosmos DB that is compatible with Azure Table storage, so moving across mostly means swapping the connection string.
- Partition key
Controls how data is spread out. Azure Cosmos DB puts items whose values match into one logical partition, and Azure Table storage stores rows sharing a PartitionKey side by side.
- Point query
An Azure Table storage lookup that supplies both the PartitionKey and the RowKey, so exactly one entity comes back; this is the fastest kind of query the service offers.
- Range query
Second only to a point query for efficiency in Azure Table storage: it names one PartitionKey plus a span of RowKey values and returns the contiguous entities within that partition.
- RowKey
An Azure Table storage property, unique per entity within one partition. Paired with the PartitionKey it gives every entity its unique key; no other properties are indexed.
- Table projection
A knowledge store projection that writes rows to Azure Table Storage, linked through generated keys (source, generatedKeyName, tableName). It works well for analysis in Power BI.