Can we join tables in on-premise SQL Server database to tables in Delta tables in Azure Delta lake? what are my options - sql

Question

I am archiving rows that are older than a year into ADLSv2 as delta tables, when there is a need to report on that data, I need to join archived data with some tables existing on on-premise database. Is there a way we can do a join without re-hydrating from or hydrating data to cloud?

Answer 1

Yes, you can achieve this task by using Azure Data Factory.
Azure Data Factory (ADF) is a fully managed, serverless data integration
service. Visually integrate data sources with more than 90 built-in,
maintenance-free connectors at no added cost. Easily construct ETL and
ELT processes code-free in an intuitive environment or write your own
code.
Firstly, you need to install the Self-hosted Integration Runtime in your local machine to access the on-premises SQL Server in ADF. To accomplish this, refer Connect to On-premises Data in Azure Data Factory with the Self-hosted Integration Runtime.
As you have archived the data in ADLS, you need to change the Access tier of that container from Cold -> Hot in order to retrieve the data in ADF.
Later, create a Linked Service using Self-hosted IR which you have created. Create a Dataset using this Linked Service to access the on-premises database.
Similarly, create a Linked Service using default Azure IR. Create a Dataset using this Linked Service to access the data from ADLS.
Now, you also require a destination database where you will store the data after join. If you are storing it in same on-premises database, you can use the existing Linked Service but you need to create a new Dataset mentioning the destination table name.
Once all this configuration done, create a Data Flow activity pipeline in ADF.
Mapping data flows are visually designed data transformations in Azure
Data Factory. Data flows allow data engineers to develop data
transformation logic without writing code. The resulting data flows
are executed as activities within Azure Data Factory pipelines that
use scaled-out Apache Spark clusters.
Learn more about Mapping data flow here.
Finally, in data-flow activity, your sources will be on-premises dataset and ADLS dataset which you have created above. You will be using join transformation in mapping data flow to combine data from two sources. The output stream will include all columns from both sources matched based on a join condition.
The sink transformation will take your destination dataset where the data will be stored as an output.