An analytic query refers to a query generating results from a semantic model. Power BI visuals, behind the scenes, send analytic queries to query the model. These queries are crafted in Data Analysis Expressions (DAX). Importantly, users aren't required to manually write DAX statements; configuring report visuals involves mapping fields from the semantic model.
Analytic queries undergo three sequential phases:
Filter: Application of filtering criteria to refine data.
Group: Grouping data based on specified parameters.
Summarize: Aggregating data to produce summarized results.
In Power BI reports, the process of filtering, akin to slicing, aims to isolate relevant data. Filters in Power BI can be applied at three levels: the entire report, a specific page, or an individual visual. Additionally, when row-level security (RLS) is enforced, filtering occurs in the background. Filters can either be inherited by or directly applied to each report visual.
Grouping, equivalent to dicing, involves segmenting query results into distinct groups.
Summarization results in a singular value. Typically, numeric columns are summarized using methods like sum or count. While these are simple summarizations, more intricate summarizations, such as a percentage of the grand total, can be achieved by defining measures in Data Analysis Expressions (DAX).
It's not mandatory for every analytic query to undergo filtering, grouping, and summarization:
Filtering is commonly applied to report visuals based on criteria like time periods or geographic locations.
Grouping is discretionary; visuals like card visuals, designed for displaying a single value, don't necessitate grouping.
Most report visuals involve summarization, though there are exceptions, like the slicer visual, which doesn't deal with summarization.
SWETA SARANGI