What's the difference between the Databricks Delta and Databricks Storage endpoints?

[MoeE]

Hi,

I see that there are two endpoints for Databricks, and I am confused as to what the difference is between these two. Can someone please explain? Thank you

Kind regards,

Mohammed

[OritA]

Hi Mohammed, 

In general: 

Databricks Lakehouse is the data management and analytics layer (where you query, govern, and process data).

While Databricks Cloud Storage is the underlying raw storage layer (where files actually live — in S3, ADLS, or GCS).

Regards,

Orit

[Nagaraju_KCS]

Analytics & Reporting - Connecting Qlik Sense to the curated, clean, and reliable data in the Databricks Lakehouse. This is the more preferred method for BI.

Data Integration & Staging - Primarily used by Qlik Data Integration tools (like Qlik Replicate) as a staging area to load raw data before it becomes a Delta Table. (e.g., Azure Data Lake Storage, Amazon S3)

[MoeE]

Hi @OritA,

Thanks for the response.

I think I am still a little confused. In Qlik Replicate, if I selected Databricks Delta as the target endpoint, then I assume that my tables and their data will be sent to and created in Databricks storage (S3, ADLS, GCS). The tables would be in the Delta table format allowing users to utilize the benefits of the Delta format.

If I selected Databricks Storage as the target endpoint in Qlik Replicate, then what is the difference? Will the data not be created in the Delta format?

Regards,

Mohammed

[MoeE]

Hi @Nagaraju_KCS , @OritA 

I am referencing these two target endpoints in Qlik Replicate:

• Databricks Lakehouse (Delta)
• Databricks (Cloud Storage)

Thank you.

Regards,

Mohammed

[john_wang · Accepted solution]

Hello Mohammed, @MoeE 

The key differences are:

1. Databricks Lakehouse (Delta)
   The data resides in cloud object storage but is governed by Delta Lake metadata, giving it full database-like behavior.
   In simple terms: cloud storage + metadata + ACID transactions = Delta tables.
   Because of this, Replicate can perform INSERT/UPDATE/DELETE operations on Delta tables just like a traditional database target.

2. Databricks (Cloud Storage)
   This refers to the raw object storage linked to your Databricks workspace.
   It’s primarily used for storing files such as CSV, JSON, Parquet, logs, images, etc.
   In Qlik Replicate, this behaves as a file-based target, not a transactional table system.

Hope this clarifies the difference.

John

[MoeE]

Hi John,

Thank you! This is a clear explanation. Have a good day.

Cheers,

Mohammed

[john_wang]

Thank you for your support Mohammed! @MoeE