Building a Data Warehouse Dimensional Model using Azure Synapse Analytics SQL Serverless

preview_player
Показать описание
The Serverless SQL Pools service within Azure Synapse Analytics allows querying CSV, JSON and Parquet data in Azure Storage, Data Lake Gen1/2 and Cosmos DB. With this functionality we are able to create a Logical Data Warehouse over data stored in these systems without moving and loading the data. However, the source data may not be in the best possible format for analytical workloads...

In this session we'll be looking at using Azure Synapse Analytics SQL Serverless Pools to create a Data Warehouse using the Dimensional Modelling technique to create a set of Dimensions and Facts and store this data in a more appropriate structure and file format.

All data will be stored in an Azure Data Lake Gen2 account with processing and serving performed by the SQL Serverless Pools engine.
Рекомендации по теме
Комментарии
Автор

13min in, already obvious this is a fantastic video. Thanks for doing this!

BaGua
Автор

i enjoyed watching this video but please get us 1080p or 4K going forward. THANKS

demdemdem-gj
Автор

How do you take care of scd 1, scd 2 etc ? Coming from traditional SQL server background, this all sounds f'ed up to me. They talk about external table, then views, then CSV files 😯 ?? W t f is going on

artus
Автор

Very helpful and clear. Thanks for sharing this.

smw
Автор

Great session - really like your "show it for real" demo style. So the use of external tables in this session is primarily to easily transform CSV to Parquet format? Also, for adding new partitions to the view, it relies on the Parquet file remaining after a Drop External table has been issued?

geehaf
Автор

Super nice !! I is there any way to have the scripts and source files ? Thank you !!

valentinloghin
Автор

Great, thanks for sharing! I was trying to implement a DataVault logical DWH but there is no hashbytes :| I hope this feature will be supported in the future

RodrigoBocanegraCruz
Автор

Would looping through a series of dates be synchronous? Is there a way to do that asynchronously?

Rothbardo
Автор

I m surprised I can't see anything

sanishthomas