topic Re: Medallion architecture, how to update Gold tables? in Data Engineering

Medallion architecture, how to update Gold tables?

Mado — Tue, 07 Mar 2023 12:30:25 GMT

Assume that I have a data source that is ingested to a few bronze tables, and transformed to a silver table. Ans next, a gold table is created by aggregating the silver table.

If new records arrive in the data source, bronze and silver tables are updated by appending new records. Since the gold table contains aggregated values, using "append" is meaningless.

I'd like to know which approach is recommended to update gold tables in case of having a large dataset:

1) Drop the current gold table, and re-create it

2) Overwrite the gold table

The first option seems to be slower if we deal with a large dataset. However, I'd like to know if there is any risk with option 2 (e.g. if the table is not overwritten correctly).

Re: Medallion architecture, how to update Gold tables?

daniel_sahal — Tue, 07 Mar 2023 14:18:46 GMT

@Mohammad Saber

Why not using MERGE? Or even CDF + Merge to do increments.

https://sarnendude.com/delta-lakes-change-data-feed-cdf-demo-in-azure-databricks/

Re: Medallion architecture, how to update Gold tables?

Mado — Tue, 07 Mar 2023 21:24:57 GMT

When there is aggregated value in the Gold table, I think it should be calculated using all records in the Silver table.

Re: Medallion architecture, how to update Gold tables?

Anonymous — Sat, 01 Apr 2023 01:10:16 GMT

Hi @Mohammad Saber

Thank you for your question! To assist you better, please take a moment to review the answer and let me know if it best fits your needs.

Please help us select the best solution by clicking on "Select As Best" if it does.

Your feedback will help us ensure that we are providing the best possible service to you. Thank you!

Re: Medallion architecture, how to update Gold tables?

Mado — Sat, 01 Apr 2023 05:47:33 GMT

Hi @Vidula Khanna

The answer didn't fit my question.

In the case of using Merge, I found a good article here:

https://medium.com/@avnishjain22/simplify-optimise-and-improve-your-data-pipelines-with-incremental-etl-on-the-lakehouse-61b279afadea