cancel
Showing results forย 
Search instead forย 
Did you mean:ย 
Data Engineering
Join discussions on data engineering best practices, architectures, and optimization strategies within the Databricks Community. Exchange insights and solutions with fellow data engineers.
cancel
Showing results forย 
Search instead forย 
Did you mean:ย 

Delta Table in DLT pipeline

habs
New Contributor II

Hi all, Is it possible to create a Delta table using DLT pipeline. I'm able to create a Delta table using Job and also able to create Materialized view and Streaming table using DLT pipleines. But I want to create Delta table using DLT pipelines. Is it possible and how can we create that?
Any help is very much appreciated, Thanks

this is how delta table looks in databricks, attaching in media

 

1 ACCEPTED SOLUTION

Accepted Solutions

iyashk-DB
Databricks Employee
Databricks Employee

Hi, yes, both streaming tables and materialized views in a DLT pipeline (now called Lakeflow Declarative Pipelines) are Delta tables under the hood. There isn't a separate "plain Delta table" object type inside a pipeline, streaming tables and materialized views are the two ways a pipeline writes Delta tables, they just differ in how the data gets refreshed.

If you specifically want a table that behaves like a plain batch table with no streaming semantics, a materialized view is the closest equivalent since it's a full batch recompute each time rather than a continuous append.

View solution in original post

1 REPLY 1

iyashk-DB
Databricks Employee
Databricks Employee

Hi, yes, both streaming tables and materialized views in a DLT pipeline (now called Lakeflow Declarative Pipelines) are Delta tables under the hood. There isn't a separate "plain Delta table" object type inside a pipeline, streaming tables and materialized views are the two ways a pipeline writes Delta tables, they just differ in how the data gets refreshed.

If you specifically want a table that behaves like a plain batch table with no streaming semantics, a materialized view is the closest equivalent since it's a full batch recompute each time rather than a continuous append.