How to use LLMs in SQL Statements with Big Query ML

Google have introduced the ability to use LLMs within SQL queries in Google Big Query with Big Query ML.
The steps to get this running are:
- Create a connection to Vertex within a Google Big Query
- Give the Connection access to Vertex AI by adding the service account to a Vertex AI user role in IAM.
- Create the model in a dataset
- Query any data, just include a required field called prompt with you can concat fields and prompts together into.
I first came across this Google Big Query LLM functionality in this medium article so shoutout to Pooja and check out her article. / bringing-gen-ai-to-the...
Here is a link to the google documentation including models that are usable in this setup: cloud.google.com/bigquery/doc...

[@user-ii1rs6pf2l]

Thank you so much for walking through the steps. The paper by Abirami Sukumaran is not exactly beginner friendly, it misses careful guidance but a video example solves this problem perfectly.

[@analyticswithadam]

No worries, it look me a while to figure out from the paper but I wouldn't have known it was possible without the paper.
Medium is really good for sparking thoughts