topic Maintaining Order Consistency: Table Creation in Databricks SQL vs. DLT Pipeline in Get Started Discussions

Maintaining Order Consistency: Table Creation in Databricks SQL vs. DLT Pipeline

Archana_Mathan — Wed, 17 Apr 2024 11:50:00 GMT

I have a CTE table with the below names as values. My objective is to create another table by concatenating all the rows from the CTE table in ascending order, resulting in the final output sequence: "Abi, Rahul, ram, Siva". When executing the query in DataBricks SQL, the expected outcome is achieved. However, when creating the table through the DLT pipeline, the order is altered, resulting in the following sequence: "Abi, Siva, ram, Rahul". Is there a solution to maintain the original order consistently, regardless of whether the table is created manually or through the DLT pipeline?

Re: Maintaining Order Consistency: Table Creation in Databricks SQL vs. DLT Pipeline

-werners- — Wed, 17 Apr 2024 11:59:46 GMT

when writing, order is not guaranteed due to the nature of distributed processing.
If you want the order to be guaranteed, you should order it when reading the data.

Your query does not write any data, DLT does, that is the difference.

Re: Maintaining Order Consistency: Table Creation in Databricks SQL vs. DLT Pipeline

standup1 — Fri, 05 Jul 2024 14:32:47 GMT

Hi @Archana_Mathan, did you solve this problem by any chance? I am having a similar problem maintaining the order of the dlt.

Thank you,