FUNCTION

In the Ultra Transcenders books

DBX-GENAI

Each book explains FUNCTION in context, with comparison tables and the common traps.

Terms in this definition

• COUNT

  A DAX function returning how many non-blank numbers, dates or text values a column contains. Boolean columns need COUNTA instead, and for counting the rows of a table COUNTROWS is the better choice.

• Unity Catalog

  Azure Databricks' governance solution covering both data and AI in one place, with centralised permissions, auditing, data discovery and lineage.

• Authorisation code

  OAuth 2.0 grant used by native and web apps: the user signs in, and the app then acts on their behalf with delegated permissions.

• REST

  Short for representational state transfer, the style of HTTP API that Azure services expose.

Related terms

• ADDCOLUMNS

  A DAX function that takes a table and hands it back with extra computed columns appended, working out each new expression row by row while keeping every existing column.

• ADX

  Azure Data Explorer is Microsoft's KQL-based service for analysing large volumes of data. Log Analytics can reach it through the adx() function, though queries and jobs in the Sentinel data lake cannot.

• ai_classify

  Sorts content under labels you define, whether plain text or the VARIANT that ai_parse_document returns. This task-specific AI Function can also take instructions and a description of each label.

• ai_extract

  An AI Function for one specific job: turning text or documents into structured fields. You describe what you want in a schema, which can be a plain list of field names or include nested objects and arrays.

• ai_mask

  A task-specific AI Function in Public Preview. It swaps the entity types you name in a piece of text for [MASKED], and its output may differ between calls.

• ai_parse_document

  Turns a BINARY document into VARIANT holding its structured content. This AI Function's output can then feed ai_prep_search, ai_classify or ai_extract.

• ai_prep_search

  Readies content for a search index: this AI Function, still in Beta, takes plain text, Markdown or the result of ai_parse_document and returns meaning-based chunks carrying context from the document.

• ai_query

  An all-purpose AI Function: it calls a Model Serving endpoint you already have from PySpark or SQL, handling batch and real-time inference, and interprets what comes back.