Tag: sql

Most classification workflows require exporting data to Python, running a model, and importing results back into your warehouse. Dremio's `AI_CLASSIFY` function eliminates that entire pipeline. You write a SELECT statement, pass in your text and your categories, and the LLM assigns a label. The classified data stays in your lakehouse, governed and queryable immediately.

Classify Your Data with SQL: A Hands-On Guide to Dremio's AI_CLASSIFY Function

Every data team has a version of this problem: a table full of raw data that needs human-readable summaries, translations, or narrative descriptions. Product descriptions that need rewriting for a new market. Customer records that need one-sentence executive summaries. Support interactions that need post-call notes.

Generate Summaries and Insights with Dremio's AI_COMPLETE Function

Unstructured text is the most underused data in most organizations. Customer emails sit in inboxes. Contract notes live in text fields. Meeting summaries exist as free-text columns in CRM systems. The information is there, but it's locked inside prose that SQL can't filter, join, or aggregate.

Extract Structured Data from Text with Dremio's AI_GENERATE Function

Intro to SQL using Apache Iceberg and Dremio