Microsoft Azure & FabricLive · 108 questions · 12 notes

PL-300: Microsoft Power BI Data Analyst

Free practice exam, study notes with architecture diagrams and the official blueprint for Microsoft Power BI Data Analyst.

Passing score
700 / 1000
Domains
4
Exam code
PL-300

Official PL-300 exam blueprint

Domains and weights from Microsoft Azure & Fabric's exam guide.

DomainWeight
Prepare the data25–30%
Model the data25–30%
Visualize and analyze the data25–30%
Manage and secure Power BI15–20%

Free PL-300 sample questions

The first questions of the free diagnostic. Pick your answer, then check it. Why every option is right or wrong, and the What-If twist, open in the workspace.

  1. Question 1 · Get or connect to data

    A marketing analyst holds only the Viewer role in the finance workspace that hosts the certified Sales semantic model. She must build her own report on that model without copying its data, and company policy forbids exporting finance data to Excel files. What does she need?
    1. ABuild permission on the Sales semantic model, then a live connection from Power BI Desktop
    2. BAn export of the model's tables to Excel to import into a new model
    3. CThe Admin role on the finance workspace
    4. DA new Import model that copies the finance database tables
    Check the answer

    Answer: A. Build permission on a shared semantic model lets people create reports on it with a live connection, without copies.

    Official docs: Microsoft Learn: Build permission for shared semantic models
    Unlock Full 4-Option Autopsy & What-If Twist in Workspace
  2. Question 2 · Design and implement a data model

    Visuals timed out and a customer count measure returned confusing totals right after a modeler set every relationship in a star-schema sales model to filter in both directions. The report must keep its current slicers and pages, and only the model may change. What should the modeler change?
    1. ASet the dimension-to-fact relationships back to a single cross-filter direction
    2. BChange every relationship to many-to-many
    3. CAdd a bi-directional relationship between the dimension tables
    4. DMark the Customer table as a date table
    Check the answer

    Answer: A. Keep relationships single-direction from dimensions to facts; use bi-directional filtering only where a scenario needs it.

    Official docs: Microsoft Learn: Bi-directional relationship guidance
    Unlock Full 4-Option Autopsy & What-If Twist in Workspace
  3. Question 3 · Get or connect to data

    A retailer's sales analysts build a Power BI report on a two-gigabyte orders table in Azure SQL Database. The data only has to be as fresh as last night's load, every visual must respond instantly, and the model must stay within a Pro workspace. Which storage mode should the orders table use?
    1. ADirectQuery, so every visual queries Azure SQL Database directly
    2. BImport, refreshed nightly after the load
    3. CDual for every table in the model
    4. DA live connection to a shared semantic model that does not exist yet
    Check the answer

    Answer: B. Import gives the fastest visuals for data that only needs to be as fresh as the last refresh; DirectQuery trades speed for currency.

    Official docs: Microsoft Learn: Table storage mode in Power BI semantic models
    Unlock Full 4-Option Autopsy & What-If Twist in Workspace
Answer all 10 free PL-300 questions

PL-300 study notes: key concepts

What each note covers. The full notes, with summaries, examples and memory hooks, open with a free account.

Architecture flow · Get or connect to data

How Power BI reaches cloud and on-premises data

Power BI Desktop author

  • leads to Semantic model in a workspace · publish over HTTPS 443

Microsoft cloud

Azure SQL Database

Power BI service

Semantic model in a workspace

  • leads to Azure SQL Database · Import or DirectQuery

Published reports

  • leads to Semantic model in a workspace

Corporate network

On-premises data gateway

  • leads to Semantic model in a workspace · outbound HTTPS 443 to the service
  • leads to On-premises SQL Server

On-premises SQL Server

Drawn fromMicrosoft Learn: What is an on-premises data gateway?

Get or connect to data

  • Import mode loads data into the VertiPaq engine, giving the fastest visuals and full DAX, but the report only shows data as fresh as the last scheduled refresh.
  • DirectQuery sends queries to the source for every visual, so data stays current, but performance depends on the source and some transformations are limited.
  • Dual storage mode lets a dimension table act as Import or DirectQuery as needed, which keeps relationships regular when facts use different storage modes.
  • A live connection to a published semantic model lets authors build reports on a shared model without copying it, and requires Build permission on that model.
  • Privacy levels decide whether Power Query can combine and fold data across sources, so a Private source is never sent to another source during evaluation.

Profile and clean the data

  • Column quality shows the percentage of valid, error and empty values in each column, so it is the quickest view for spotting cleaning work across a table.
  • Column distribution counts distinct and unique values, which tells you whether a column could act as a key before you build relationships on it.
  • Profiling uses only the first 1,000 rows by default, so switch to the entire data set before trusting counts on a large table.
  • Step-level errors stop the query, such as a renamed source column, while cell-level errors affect single values and can be replaced or removed.
  • Power Query comparisons are case-sensitive, so values that differ only in letter case survive duplicate removal unless you transform their case first.

Transform and load the data

  • Unpivot turns columns such as Jan, Feb and Mar into attribute and value rows, which creates the tall table shape that a star schema fact table needs.
  • Merge joins two queries on matching columns with a chosen join kind, while append stacks queries with the same columns into one longer table.
  • A left anti join returns only the rows in the first table that have no match in the second, which is ideal for finding orphan records.
  • Query folding sends transformation steps to the source as one native query, so keep foldable steps early and check where step folding indicators stop.
  • Turning off Enable load on staging queries keeps them out of the model while reference queries can still build fact and dimension tables from them.

Design and implement a data model

  • A star schema keeps facts with numbers at a stated grain and describes them with flattened dimension tables, which gives simpler DAX and faster queries.
  • One-to-many relationships with single direction let filters flow from each dimension to the fact table, which is the default and safest model design.
  • Bi-directional cross-filtering can create ambiguous filter paths and slower queries, so use it only where needed, such as a bridge table in a many-to-many design.
  • Only one relationship between two tables can be active, so a role-playing Date table uses inactive relationships that USERELATIONSHIP activates inside a measure.
  • A date table needs one row for every day without gaps across the fact dates and should be marked as a date table for time intelligence.

Create model calculations by using DAX

  • A CALCULATE filter argument replaces existing filters on the same column, KEEPFILTERS intersects them, and REMOVEFILTERS clears them, which drives most filter context questions.
  • Calculated columns have a row context but no filter context, so an aggregation inside one needs CALCULATE to transition the row into a filter.
  • Time intelligence functions need a contiguous, marked date table and accept a year-end argument such as 6/30 for fiscal years ending in June.
  • Semi-additive values such as stock on hand are summed across products but not across time, so use a closing balance function for each period.
  • Calculation groups wrap SELECTEDMEASURE in shared logic such as year-to-date, so a few items replace dozens of near-identical explicit measures.

Optimize model performance

  • Performance Analyzer splits each visual's load time into DAX query, visual display and other time, which tells you whether to fix measures or the page design.
  • Removing unused columns and summarizing facts to the grain reports use are the most effective ways to shrink an Import model and speed up refresh.
  • VertiPaq stores numeric columns more efficiently than text, so converting unique text identifiers into whole numbers reduces model size noticeably.
  • Simple row calculations belong in Power Query or the source, because DAX calculated columns compress less well and are built after every table loads.
  • Boolean filter arguments such as KEEPFILTERS on one column let the column store filter efficiently, while FILTER over a whole table evaluates row by row.
Read the full PL-300 study notes

PL-300 FAQ

What is the passing score for PL-300?

The passing score for Microsoft Power BI Data Analyst is 700 / 1000.

Which domains does PL-300 cover?

The official exam guide lists 4 domains: Prepare the data (25–30%); Model the data (25–30%); Visualize and analyze the data (25–30%); Manage and secure Power BI (15–20%).

Is there a free PL-300 practice test?

Yes. Gyan AI's free diagnostic is 10 PL-300 questions with no sign-up, each with an explanation for every option and a link to the official documentation. A free account unlocks the full question set and study notes.

Are these the real PL-300 exam questions?

No. Real exam content is confidential under vendor exam agreements, and using leaked "dumps" can get a certification revoked. Gyan AI questions are original and written to the published skills outline.