PL-300 Study Guide

Power BI connects to hundreds of sources through Get data, from files and folders to relational databases, online services and OData feeds. Instead of importing your own copy, you can also connect to an existing shared semantic model so many reports build on one governed dataset.

What you need to know

• Get data groups sources into File, Database, Power Platform, Azure, Online Services and Other; Recent sources and OneLake data hub speed up common connections
• A live connection to a Power BI semantic model or Analysis Services means the model lives elsewhere; you build visuals but cannot add tables or change the model
• Connecting to a shared semantic model promotes reuse and a single version of the truth across many reports
• Some connectors support query folding, which pushes your transformations back to the source for speed

Microsoft Learn resource: Data sources in Power BI Desktop (opens in a new tab)

Data source settings let you update where a query points and how it authenticates without rebuilding the query. Privacy levels tell Power Query whether data from two sources may be combined, which protects sensitive data from leaking into a public source.

What you need to know

• Data source settings are reached from Transform data, and cover the sign-in credentials, the connection details and the privacy level per source
• Credentials come in several kinds depending on the source: Windows, database username and password, Microsoft account, OAuth or a key
• The three privacy levels are Private, Organizational and Public; Private is the most restrictive and never flows to another source
• A privacy level mismatch, or setting Ignore privacy levels, can change results or break query folding

Microsoft Learn resource: Power Query privacy levels (opens in a new tab)

Exam tip: If two sources will be merged, their privacy levels decide whether the combine is even allowed, so know that Private data is never sent to a Public source.

Storage mode decides where your data actually sits and when it is read. Import caches a compressed copy in the model, DirectQuery leaves the data at the source and queries it live, and Direct Lake reads Fabric OneLake data directly for near-import speed without a refresh.

What you need to know

• Import is the default and usually the best choice for performance
• DirectQuery suits very large or highly volatile data, or where data cannot be copied
• Composite models let one model mix Import and DirectQuery tables
• Direct Lake is a Microsoft Fabric feature that reads Delta tables in OneLake

Microsoft Learn resource: Storage mode in Power BI Desktop (opens in a new tab)

Parameters store a value you can reuse and change in one place, such as a server name, a file path or a filter value. They make a query portable across environments and drive features like incremental refresh.

What you need to know

• Parameters are created and edited in Manage Parameters in the Power Query Editor
• A parameter has a type and can take any value, a list of suggested values, or a query's output
• Parameters can feed a data source path so you switch from dev to production by changing one value
• Templates (.pbit) prompt for parameter values when opened, which is how you distribute a report that connects to each user's own source

Microsoft Learn resource: Using parameters in Power Query (opens in a new tab)

Before you trust data, you profile it. Power Query's data profiling tools show the shape and quality of each column so you can spot problems early, right inside the editor.

What you need to know

• Column quality shows the percentage of valid, error and empty values in a column
• Column distribution shows how many distinct and unique values a column holds
• Column profile gives per-column statistics (min, max, average, counts) and a value distribution chart
• By default profiling is based on the top 1,000 rows; you can switch it to the entire data set at the bottom of the window

Microsoft Learn resource: Data profiling tools in Power Query (opens in a new tab)

Exam tip: Remember that profiling defaults to the top 1,000 rows; switch to the whole data set when a quality problem might hide further down.

Real data is messy: inconsistent casing, stray whitespace, wrong or missing values. Power Query gives you targeted transformations to standardize and repair columns so downstream calculations are reliable.

What you need to know

• Replace values swaps a specific value for another; Replace errors turns error cells into a chosen value such as null or zero
• Fill Down and Fill Up copy a value into adjacent null cells, useful after unpivoting or with merged-cell sources
• Format transforms handle Trim, Clean, and casing (lowercase, UPPERCASE, Capitalize Each Word)
• Remove Duplicates, Remove Blank Rows and filtering out unwanted values clean the row set
• Keep the source data unchanged: every fix is a recorded, repeatable Applied Step

Microsoft Learn resource: Clean, transform, and load data in Power BI (opens in a new tab)

Import errors happen when a value cannot fit a column's data type, when a source changes, or when a step references something that no longer exists. Knowing where errors surface and how to handle them keeps a refresh from failing.

What you need to know

• Cell-level (value) errors come from failed type conversions and show as Error in the cell; step-level errors stop the whole query
• Keep Errors and Remove Errors let you isolate or discard bad rows; Replace Errors substitutes a safe value
• A common cause is setting a data type too early, before a text value is cleaned
• Renamed or deleted source columns break steps that reference them by name

Microsoft Learn resource: Dealing with errors in Power Query (opens in a new tab)

Every column has a data type, and it drives how values sort, aggregate and behave in DAX. Setting the right type early prevents subtle errors and keeps model size down.

What you need to know

• Common types include Decimal number, Fixed decimal number (currency), Whole number, Date, Date/time, Text, True/false and Binary
• Fixed decimal number stores four decimal places and avoids the rounding surprises of floating point, so it suits money
• The type icon sits in each column header; setting a type adds a Changed Type step
• Use Using Locale when a date or number is written in a different regional format

Microsoft Learn resource: Data types in Power Query (opens in a new tab)

You reshape columns to fit the model: splitting, extracting, deriving and formatting. Power Query offers ready-made transforms plus custom columns for anything the ribbon does not cover.

What you need to know

• Custom Column builds a new column from an M expression using other columns
• Column from Examples infers the transformation from a few values you type
• Conditional Column creates if/then/else logic without writing M
• Split Column (by delimiter, number of characters or position) and Extract (before, after, between) break text apart
• Prefer a Power Query column for anything that can be computed at load; save DAX calculated columns for model logic

Microsoft Learn resource: Add a custom column in Power Query (opens in a new tab)

Group By collapses rows that share a value into summary rows, computing counts, sums, averages and more. It is how you change the grain of a query, for example from one row per sale to one row per product.

What you need to know

• Group By supports Sum, Count, Count Distinct, Average, Min, Max and All Rows
• Basic grouping uses one column; Advanced grouping adds more columns and more aggregations at once
• The All Rows aggregation keeps the underlying rows as a nested table for later expansion
• Grouping reduces row count and can shrink the model, but it also changes the grain, so do it deliberately

Microsoft Learn resource: Group or summarize rows in Power Query (opens in a new tab)

Pivoting and unpivoting move data between wide and tall shapes. For analytics you almost always want tall (unpivoted) data, because a column per category is hard to model.

What you need to know

• Unpivot turns columns into attribute/value pairs (tall data), ideal for a model; Unpivot Other Columns keeps chosen columns fixed
• Pivot does the reverse, turning row values into columns, which you rarely want for a model
• Transpose flips rows and columns wholesale, handy when headers arrive as a row
• Unpivot Other Columns is the resilient choice because it still works when new period columns are added later

Microsoft Learn resource: Unpivot columns in Power Query (opens in a new tab)

Exam tip: When a source has one column per month and new months keep appearing, Unpivot Other Columns survives the change while a fixed Unpivot does not.

Sources like JSON, XML and nested API responses arrive as records and lists rather than a flat table. Power Query expands these structures into rows and columns you can work with.

What you need to know

• A Record is a single set of fields; a List is a set of values; both appear as clickable links in a cell
• Expand adds a record's fields as new columns; Expand to New Rows turns a list into rows
• The JSON and XML connectors return nested structures you drill into with these expand steps
• Convert a List to a table with To Table before you can transform it further

Microsoft Learn resource: Import data from a JSON file (opens in a new tab)

A well-shaped model separates facts (the events you measure) from dimensions (the things you slice by). Building these tables in Power Query sets you up for a clean star schema in the model.

What you need to know

• A fact table holds measurements and foreign keys, is long and narrow, and grows over time (Sales, Orders)
• A dimension table describes an entity, is shorter and wider, and holds the attributes you filter and group by (Product, Customer, Date)
• Reference queries let you build several dimensions from one source query
• Duplicate and Remove Duplicates on a key column is a common way to carve a dimension out of a fact source

Microsoft Learn resource: Understand star schema and its importance for Power BI (opens in a new tab)

Duplicate and Reference both spin a new query off an existing one, but they behave very differently, and the difference affects refresh and maintenance.

What you need to know

• Duplicate copies every step into an independent query; later edits to the original do not affect the copy
• Reference creates a new query that starts from the output of the original, so it inherits later changes
• Reference is the tool for building several tables from one cleaned base query (staging queries)
• A referenced source may be evaluated more than once at refresh unless the source query is loaded or folding applies

Microsoft Learn resource: Reference queries in Power Query (opens in a new tab)

Merge and Append are the two ways to combine queries, and they answer different questions: Merge adds columns, Append adds rows.

What you need to know

• Merge offers join kinds: Left Outer, Right Outer, Full Outer, Inner, Left Anti and Right Anti
• Fuzzy matching in Merge pairs near matches when keys are not identical
• Append works cleanly when columns share names; mismatched names create extra columns

Microsoft Learn resource: Merge queries overview (opens in a new tab)

Exam tip: Merge adds columns (a join), Append adds rows (a union). Watch the wording of the scenario to pick the right one.

Relationships need a matching column on each side, and it works best when one side has unique values. Preparing clean keys in Power Query avoids ambiguous or broken relationships in the model.

What you need to know

• A dimension's key should be unique with no blanks; the fact side holds the matching foreign key
• When no single column is unique, create a composite key by merging columns (for example Region plus Store)
• Keys must share a data type on both sides for the relationship to work
• Blank or duplicate keys on the one side cause relationship errors or an unexpected many-to-many

Microsoft Learn resource: Create and manage relationships in Power BI (opens in a new tab)

Not every query should load into the model. Controlling load and refresh behavior keeps the model lean and refreshes efficient.

What you need to know

• Enable load controls whether a query becomes a model table; turn it off for staging queries used only as references
• Include in report refresh controls whether a query refreshes with the report
• A query with Enable load off still exists and can be referenced by other queries
• Disabling load on intermediate queries reduces model size and clutter in the Fields pane

Microsoft Learn resource: Loading data in Power Query (opens in a new tab)

Beyond names and data types, tables and columns carry properties that change how they behave in reports: how they sort, how they summarize, and how visuals treat them. Setting these once in the model saves work in every report built on it.

What you need to know

• Sort by column makes one column sort by another, most often a month name sorted by a month number
• Data category (Address, City, Web URL, Image URL, and so on) tells visuals how to treat a column, which drives maps and image display
• Default summarization sets whether a numeric column sums, averages or does not aggregate; set numeric keys to Do not summarize
• Hiding technical columns (keys, helper columns) keeps the Fields pane clean for report authors

Microsoft Learn resource: Specify data categories in Power BI Desktop (opens in a new tab)

A role-playing dimension is one dimension used several times for different roles, the classic case being a Date table joined to Order Date, Ship Date and Due Date. You keep one physical table and switch which relationship is active in a calculation.

What you need to know

• Only one relationship between two tables can be active at a time; the others are inactive (dashed line)
• USERELATIONSHIP activates an inactive relationship for a specific measure
• The alternative is to import the dimension more than once (one table per role), which uses more model space
• Role-playing keeps a single date table and a single set of attributes, which is cleaner to maintain

Microsoft Learn resource: Active vs inactive relationship guidance (opens in a new tab)

Every relationship has a cardinality (how the keys match up) and a cross-filter direction (which way filters travel). Getting these right is the single biggest driver of correct results.

The four cardinalities:

1. One-to-many (1:*): the most common; one row on the dimension side matches many on the fact side
2. Many-to-one (*:1): the same thing viewed from the fact side
3. One-to-one (1:1): each row matches exactly one row on the other side
4. Many-to-many (:): neither side is unique; use with care, it can produce ambiguous results

• Cross-filter direction is Single (filter flows from the one side to the many) or Both (bidirectional)
• Single direction is the default and the safe choice; Both can create ambiguity and performance cost
• Bidirectional filtering is sometimes needed for many-to-many scenarios and certain slicer patterns

Microsoft Learn resource: Create and manage relationships in Power BI Desktop (opens in a new tab)

Exam tip: Default to one-to-many with single cross-filter direction; reach for Both or many-to-many only when a scenario truly needs it, and know the ambiguity risk it brings.

Time intelligence in DAX needs a proper date table: one row per day, a contiguous range with no gaps, marked as the model's date table. Reusing one date table across every date relationship keeps calculations consistent.

What you need to know

• A date table has one continuous row per date covering every date used in the data, with no missing days
• Mark as date table tells Power BI which table to use for time intelligence and requires a unique date column
• You can build one with CALENDAR or CALENDARAUTO in DAX, in Power Query, or from a source
• CALENDARAUTO scans the model for the min and max dates automatically
• Add columns like Year, Quarter, Month and Month Number so visuals sort and group correctly

Microsoft Learn resource: Create date tables in Power BI Desktop (opens in a new tab)

Calculated columns and tables are computed with DAX after the data loads. They have their place, but they are not always the right tool, and the exam tests whether you know when to use each.

What you need to know

• A calculated column is evaluated row by row at refresh and stored in the model, so it costs memory
• Prefer a Power Query custom column over a calculated column when the value can be computed at load
• A measure is usually better than a calculated column when you need an aggregation that responds to filters
• Calculated tables suit model-only helpers: a date table, a role-playing copy, or a bridge table
• Calculated columns can be used as slicers, on an axis, or in a relationship; measures cannot

Microsoft Learn resource: Use calculated columns in Power BI Desktop (opens in a new tab)

Measures are DAX calculations evaluated in the filter context of a visual. A single aggregation measure wraps one aggregation function, and it is the workhorse of most reports.

What you need to know

• SUM, AVERAGE, MIN, MAX and COUNT aggregate a single column; COUNTROWS counts rows of a table
• DISTINCTCOUNT counts unique values, common for counting customers or orders
• A measure recalculates for every cell based on the filters that apply there (filter context)
• Measures do not consume storage the way calculated columns do; they compute at query time
• Write measures explicitly rather than dragging a column into a visual, so the logic is reusable and named

Microsoft Learn resource: Create measures for data analysis in Power BI Desktop (opens in a new tab)

CALCULATE is the most important function in DAX because it can change the filter context in which a measure is evaluated. Almost every non-trivial measure uses it.

What you need to know

• CALCULATE(expression, filter1, filter2, ...) evaluates the expression after applying the filters you give it
• Its filter arguments can add, replace or remove filters on the current context
• REMOVEFILTERS (or ALL) inside CALCULATE clears filters, which is how you compute a percentage of total
• KEEPFILTERS preserves existing filters instead of overwriting them
• CALCULATE also turns a row context into a filter context (context transition), which matters inside iterators

Microsoft Learn resource: CALCULATE function (DAX) (opens in a new tab)

Exam tip: When a measure needs to ignore a slicer or compute a share of total, that is CALCULATE with ALL or REMOVEFILTERS.

Time intelligence answers the questions every business asks: year-to-date, last year, and change over time. DAX has a family of functions for it, and they all rely on a marked date table.

What you need to know

• TOTALYTD, TOTALQTD and TOTALMTD give running totals within a year, quarter or month
• SAMEPERIODLASTYEAR and DATEADD shift a period for prior-period comparison
• DATESYTD and DATESINPERIOD return date tables you use inside CALCULATE
• These functions require a continuous date table marked as the date table
• Year-over-year change is typically a measure minus its SAMEPERIODLASTYEAR value, divided by that value

Microsoft Learn resource: Time intelligence functions (DAX) (opens in a new tab)

DAX includes statistical functions for the everyday math a report needs beyond a plain sum, from averages and counts to rankings.

What you need to know

• AVERAGE, MEDIAN, MIN and MAX summarize a column; the X versions (AVERAGEX, MINX, MAXX, SUMX) iterate an expression row by row
• COUNT, COUNTA, COUNTROWS and DISTINCTCOUNT count values or rows in different ways
• RANKX ranks rows by an expression and is common for top-N and league tables
• DIVIDE performs safe division, returning a chosen value (blank by default) instead of an error when dividing by zero

Microsoft Learn resource: Statistical functions (DAX) (opens in a new tab)

Exam tip: Use DIVIDE instead of the / operator so a divide-by-zero returns blank rather than an error.

Some values should not be summed across time: an account balance or an inventory count at month end is a point-in-time figure. These are semi-additive, and they need special handling over the date dimension.

What you need to know

• Additive measures sum across every dimension (sales); non-additive never sum meaningfully (a ratio); semi-additive sum across some dimensions but not time
• Typical semi-additive values are stock levels, headcount and account balances
• LASTNONBLANKVALUE, CLOSINGBALANCEMONTH and OPENINGBALANCEMONTH return the right point-in-time value
• The pattern is usually the last (or first) value in the period rather than a sum of the period

Microsoft Learn resource: Create DAX calculations in Power BI Desktop (opens in a new tab)

Quick measures build common DAX for you from a dialog, then leave the generated formula so you can learn from it or tweak it. They are a fast, error-free way to get standard calculations.

What you need to know

• Quick measures cover categories like totals, averages per category, running and moving aggregates, time intelligence, and running totals
• The dialog produces real DAX you can read and edit afterward
• Some quick measures can be generated with Copilot or from a natural-language prompt
• They are a good learning tool for seeing how CALCULATE and iterators are used in practice

Microsoft Learn resource: Use quick measures for common calculations (opens in a new tab)

Calculated tables and columns are defined with DAX and materialized in the model. They solve modeling problems that Power Query cannot, because they can reference the whole model and other DAX.

What you need to know

• A calculated table is created from a DAX table expression such as CALENDAR, FILTER, SUMMARIZE, UNION or DISTINCT
• Common uses are a date table, a role-playing dimension copy, or a bridge table for many-to-many
• Calculated tables and columns refresh with the model and add to its size
• They can reference measures and other tables, which Power Query cannot do

Microsoft Learn resource: Create calculated tables in Power BI Desktop (opens in a new tab)

Calculation groups apply one set of calculations (like time intelligence) across many measures without you writing the same variations for each. They cut down measure sprawl dramatically.

What you need to know

• A calculation group holds calculation items, each a DAX expression that uses SELECTEDMEASURE()
• One time-intelligence calculation group (YTD, PY, YoY %) can apply to every measure in the model
• They are created in the model view (or in external tools like Tabular Editor) and appear as a table with a single column
• Calculation item order matters and is controlled by an ordinal property
• They reduce the number of explicit measures you must build and maintain

Microsoft Learn resource: Calculation groups in Power BI (opens in a new tab)

The fastest model is a small model. Power BI's engine compresses columns, so removing what you do not need shrinks the file and speeds up every query.

What you need to know

• Columns cost more than rows: a high-cardinality column (like a unique ID or a datetime) is expensive to compress
• Remove columns you do not report on, and split datetime into separate date and time columns to lower cardinality
• Filter out rows you never analyze (old years, irrelevant categories) at the query level
• Disable Auto date/time to stop hidden per-table date tables from bloating the model
• A star schema with slim dimensions compresses better than a wide flat table

Microsoft Learn resource: Data reduction techniques for Import modeling (opens in a new tab)

When a report is slow, you need to find the culprit rather than guess. Performance Analyzer times each visual, and DAX query view lets you test and tune the queries behind them.

What you need to know

• Performance Analyzer records how long each visual takes, split into DAX query, visual display and other
• A long DAX query time points at the measure; a long visual display time points at the visual or too much data
• Copy query from Performance Analyzer gives you the exact DAX a visual runs
• DAX query view (in Power BI Desktop) runs DAX queries against the model to test and refine measures
• Bidirectional relationships, complex measures and too many visuals on a page are common slow points

Microsoft Learn resource: Use Performance Analyzer to examine report performance (opens in a new tab)

Granularity is how detailed the data is. Storing data at a finer grain than you ever report on wastes memory and slows queries, so pre-aggregating can help.

What you need to know

• Grain is the level of detail of one row (per transaction, per day, per month)
• Group and aggregate in Power Query, or use a summary table, when reports never need the row-level detail
• Aggregations tables can sit over a DirectQuery model so common queries hit a small cached summary
• Lower granularity means fewer rows, smaller model and faster visuals, at the cost of drill-down detail

Microsoft Learn resource: Data reduction techniques for Import modeling (opens in a new tab)

Choosing the right visual for the question is a core analyst skill, and the exam tests it directly. Each visual answers a different kind of question, so matching the visual to the data relationship matters.

What you need to know

• Comparison across categories: bar and column charts
• Trend over time: line and area charts
• Part-to-whole: pie, donut, treemap, and stacked charts (use sparingly)
• Relationship between measures: scatter chart
• A single number against a target: card, KPI, and gauge
• Detail and cross-tab: table and matrix

Microsoft Learn resource: Visualization types in Power BI (opens in a new tab)

Exam tip: A scatter chart is the go-to for showing the relationship between two (or three) numeric measures; a line chart is for a measure over time.

The Format pane controls how a visual looks and behaves, from colors and labels to axis and header settings. Consistent formatting is what separates a polished report from a rough one.

What you need to know

• The Format (paintbrush) pane splits into Visual settings and General settings (position, title, background, border)
• Data labels, legend, axis titles, and gridlines are all toggled and styled here
• The X and Y axis can be set to a fixed range, log scale, or a custom start and end
• Conditional formatting can drive colors, and settings can be copied between visuals with Format painter

Microsoft Learn resource: Visualizations in Power BI reports (opens in a new tab)

The narrative (smart narrative) visual writes a plain-language summary of what a report or visual shows, and Copilot can generate and update it for you. It turns numbers into sentences that update as filters change.

What you need to know

• The narrative visual produces dynamic text that responds to filters and slicers
• Copilot can generate the summary and let you add your own prompts for specific points
• You can insert dynamic values into the text so figures stay current
• It is useful for executive summaries and accessibility, giving a written read-out of the data

Microsoft Learn resource: Create a narrative summary with Copilot (opens in a new tab)

A report theme applies a consistent set of colors and default formatting across every visual in one place. It is the fastest way to brand a report and keep it uniform.

What you need to know

• Themes are applied from View > Themes; Power BI ships built-in themes and accepts custom ones
• A custom theme is a JSON file defining the color palette, fonts, and default visual formatting
• Customize current theme edits colors and text without leaving Power BI Desktop
• Theme colors flow into the visual color pickers as the report's palette

Microsoft Learn resource: Use report themes in Power BI Desktop (opens in a new tab)

Conditional formatting changes a visual's colors or adds icons based on the data itself, so values that need attention stand out automatically.

What you need to know

• In tables and matrices you can format background color, font color, data bars, icons, and web URLs
• Rules can be based on a color scale (gradient), explicit rules (ranges), or a field value
• The field driving the format can differ from the field being displayed
• Conditional formatting also applies to column and bar chart colors and to card values

Microsoft Learn resource: Apply conditional formatting in tables (opens in a new tab)

Slicers and the Filters pane both narrow the data a report shows, but they work differently and suit different needs. Knowing the filter scopes is heavily tested.

What you need to know

• A slicer is an on-canvas visual users click; the Filters pane is a side panel with more filter types
• Filter scopes are visual-level, page-level, and report-level, from narrowest to broadest
• Filter types include basic, advanced (conditions), and Top N
• Drillthrough filters and the sync of slicers across pages extend filtering further
• The Filters pane can be hidden, locked, or reordered for report consumers

Microsoft Learn resource: Add a filter to a report in Power BI (opens in a new tab)

Copilot can build a report page for you from a prompt, choosing visuals that answer the question you describe. It is a fast starting point you then refine.

What you need to know

• Copilot can generate a full page of visuals based on a natural-language request or the model's content
• It uses the semantic model's tables, fields, and measures to decide what to show
• Generated pages are editable like any other, so you tune the visuals afterward
• Copilot in Power BI generally requires a paid capacity (Fabric or Premium) to be enabled

Microsoft Learn resource: Create and edit Power BI reports with Copilot (opens in a new tab)

Beyond building a whole page, Copilot can suggest what to put on a page, proposing topics and visuals when you are not sure where to start.

What you need to know

• Copilot can suggest report topics and the fields or measures that best answer them
• Suggestions are grounded in the semantic model, so a well-described model gives better results
• You accept, edit, or ignore suggestions; nothing is final until you place it
• Clear table, column, and measure names help Copilot suggest relevant content

Microsoft Learn resource: Create and edit Power BI reports with Copilot (opens in a new tab)

Page-level settings control size, background, and behavior. They matter for both desktop and embedded or mobile scenarios.

What you need to know

• Page size presets include 16:9 and 4:3, plus Letter, Tooltip, and Custom dimensions
• Canvas background and wallpaper are set per page; page alignment can be centered
• Page information settings name the page and can enable it as a Q&A or tooltip page
• Hidden pages are not shown to consumers but can still be drillthrough or tooltip targets

Microsoft Learn resource: Change report page display settings in Power BI (opens in a new tab)

Paginated reports are the pixel-perfect, print-ready format built with Power BI Report Builder. They serve a different need from interactive reports, and knowing when to choose them is tested.

What you need to know

• Paginated reports are designed to print or export cleanly across many pages (invoices, statements, operational lists)
• They can render every row of a large table, unlike an interactive visual that summarizes
• They are built in Power BI Report Builder, not Power BI Desktop, and use RDL
• They support precise layout, headers and footers, and export to PDF, Excel, and more

Microsoft Learn resource: What are paginated reports in Power BI? (opens in a new tab)

Exam tip: When the requirement is a print-perfect, multi-page document that shows every row, that is a paginated report, not a standard Power BI report.

Visual calculations are DAX added directly to a visual, computed on the visual's result rather than stored in the model. They make once-hard calculations like running totals and moving averages simple.

What you need to know

• A visual calculation operates on the data already in the visual, so it can reference other rows in that visual
• Functions like RUNNINGSUM, MOVINGAVERAGE, and PREVIOUS work on the visual's axis
• Visual calculations are simpler to write than the equivalent model measures and add no model size
• They are edited in a dedicated visual calculations edit mode on the visual

Microsoft Learn resource: Using visual calculations in Power BI (opens in a new tab)

Bookmarks capture the current state of a page (filters, slicers, visibility) so you can return to it or build guided, interactive experiences.

What you need to know

• A bookmark saves filter and slicer state, sort order, and which visuals are visible
• Bookmarks can be Data, Display, and Current page scoped, and can apply to selected visuals only
• Chaining bookmarks with buttons builds toggles, story steps, and pop-up panels
• Personal bookmarks let consumers save their own views in the Power BI service

Microsoft Learn resource: Create bookmarks in Power BI Desktop (opens in a new tab)

A report-page tooltip shows a mini report page when a user hovers a visual, giving rich context beyond the default value read-out.

What you need to know

• Create a page, set its type to Tooltip, and size it as a tooltip in Page information
• Assign the tooltip page to a visual in the visual's Tooltip formatting setting
• Tooltip pages can hold their own visuals filtered to the hovered data point
• Modern tooltips can also show visuals and drill actions without a dedicated page

Microsoft Learn resource: Create tooltips based on report pages (opens in a new tab)

By default, selecting a value in one visual filters the others. Edit interactions lets you control exactly how each visual reacts, which is key to a clean report.

What you need to know

• The three interaction behaviors are Filter, Highlight, and None
• Edit interactions (on the Format ribbon) shows an icon on every other visual so you set each one
• Highlight dims non-matching data; Filter removes it; None leaves the target unchanged
• Setting interactions thoughtfully prevents a busy page from filtering itself into confusion

Microsoft Learn resource: Change how visuals interact in a report (opens in a new tab)

Buttons, page navigators, and bookmarks turn a report into an app-like experience with menus and drill paths instead of raw tabs.

What you need to know

• Buttons support actions: Back, Page navigation, Bookmark, Drillthrough, Q&A, and Web URL
• The Page navigator button auto-builds a menu of report pages that stays in sync as pages change
• The Bookmark navigator builds a set of buttons from a bookmark group
• Navigation elements make hidden and drillthrough pages reachable in a controlled way

Microsoft Learn resource: Create navigation buttons in Power BI (opens in a new tab)

How a visual is sorted changes the story it tells. Power BI sorts by a field, and you can control direction and even sort by a field that is not displayed.

What you need to know

• Use More options (...) > Sort axis (or Sort by) to choose the field and direction
• A visual can sort by a column that is not shown in the visual
• Sort by column at the model level fixes categorical sort order, such as months in calendar order
• Sorting set in Desktop is saved with the report; consumers can re-sort in the service

Microsoft Learn resource: Change how a chart is sorted in a report (opens in a new tab)

The Sync slicers pane lets one slicer control several pages, so a filter chosen on one page carries to others without duplicate slicers.

What you need to know

• The Sync slicers pane (View menu) shows, per page, whether a slicer is Visible and Synced
• A slicer can be synced to pages where it is not visible, filtering them silently
• Groups let you sync a subset of slicers together
• This keeps a consistent filter context across a multi-page report

Microsoft Learn resource: Sync and use slicers on other pages (opens in a new tab)

The Selection pane lists every object on a page and controls its visibility and stacking order. It is essential for bookmarks and layered designs.

What you need to know

• The Selection pane shows tab order, visibility (the eye icon), and layer order of every object
• Grouping objects lets you move, show, and hide them together
• Layer order (bring forward, send backward) controls what sits on top when objects overlap
• Combined with bookmarks, the Selection pane builds show/hide panels and toggles

Microsoft Learn resource: Group visuals in a Power BI report (opens in a new tab)

Drillthrough sends a user from a summary visual to a detail page filtered to the item they chose. It is the standard pattern for summary-to-detail exploration.

What you need to know

• Add a field to the Drillthrough well of a target page to make it a drillthrough destination
• Right-clicking a data point offers Drill through to the target, carrying that item's filter
• A back button is added automatically and returns the user to where they came from
• Cross-report drillthrough can jump to a page in a different report in the same workspace
• Keep all filters passes the source page's filters along with the drillthrough

Microsoft Learn resource: Set up drillthrough in Power BI reports (opens in a new tab)

Report and semantic model settings control whether and how consumers can export the underlying data, which matters for governance.

What you need to know

• From a visual, Export data can return summarized data or, if allowed, underlying data
• Export options are governed by semantic model and tenant settings; admins can restrict them
• Reports can be exported to PDF and PowerPoint from the service
• Restricting underlying-data export protects row-level detail while still allowing summaries

Microsoft Learn resource: Export data from a Power BI visualization (opens in a new tab)

The mobile layout view lets you build a phone-optimized version of a page so the report is usable on a small screen without redesigning the data.

What you need to know

• Mobile layout is a separate canvas per page where you drag in and arrange existing visuals for portrait phones
• Only visuals already on the page are available; the mobile layout reuses them
• If no mobile layout exists, the phone shows the desktop layout, which is often hard to use
• The Power BI mobile apps render the mobile layout automatically

Microsoft Learn resource: Mobile-optimized Power BI reports overview (opens in a new tab)

Personalize visuals lets consumers change a visual (its fields, visual type, and more) for their own view without edit rights, and without changing the report for others.

What you need to know

• Turn on Personalize visuals in report settings (File > Options or the service report settings)
• Consumers can swap the visual type, change measures and dimensions, and adjust the legend for their own session
• Personal changes can be saved as a personal bookmark; they never affect the published report
• It reduces the need for many near-identical report variants

Microsoft Learn resource: Let users personalize visuals in a report (opens in a new tab)

Accessible reports work for everyone, including keyboard and screen-reader users. Power BI has built-in features and design practices the exam expects you to know.

What you need to know

• Set tab order in the Selection pane so keyboard navigation follows a logical path
• Add alt text to visuals so screen readers can describe them
• Do not rely on color alone; use labels, markers, and sufficient contrast
• Themes include accessible color palettes; the narrative visual gives a text read-out
• Report readers can navigate visuals and access a data table view with the keyboard

Microsoft Learn resource: Design Power BI reports for accessibility (opens in a new tab)

Automatic page refresh keeps a page's visuals current for real-time monitoring, without a manual refresh, on supported sources.

What you need to know

• Automatic page refresh works on DirectQuery sources (and change detection), not on Import
• Fixed interval refreshes every set period; Change detection refreshes when a measure's value changes
• Minimum intervals and availability depend on capacity (Premium or Fabric allows shorter intervals)
• It is configured per page in the page formatting settings

Microsoft Learn resource: Automatic page refresh in Power BI (opens in a new tab)

Power BI can explain changes in your data automatically. The Analyze feature runs statistical analysis on a data point so you do not have to slice manually.

What you need to know

• Right-click a data point and choose Analyze to Explain the increase or Explain the decrease
• Find where the distribution is different surfaces categories that behave unlike the rest
• Power BI generates ready-made visuals explaining the change, which you can pin to the report
• It works on visuals with a measure and a category or time axis

Microsoft Learn resource: Apply Insights in Power BI to explain fluctuations (opens in a new tab)

Grouping, binning, and clustering all collapse detail into meaningful buckets, but for different data. They help turn granular values into readable categories.

What you need to know

• Grouping combines chosen categorical values into a single group (for example, several small regions into Other)
• Binning buckets a numeric or date field into equal-size ranges (bins), such as age ranges
• Clustering uses a machine-learning grouping on a scatter chart to find natural clusters
• Groups and bins appear as new fields you can use like any other

Microsoft Learn resource: Use grouping and binning in Power BI Desktop (opens in a new tab)

Power BI ships AI-powered visuals that surface insights without DAX: the Key influencers, Decomposition tree, and Q&A visuals each answer a different analytical question.

Key influencers: shows what factors most affect a chosen metric, ranking the drivers behind an outcome.

Decomposition tree: breaks a measure down across dimensions interactively, and can auto-pick the next most significant split (AI splits: High value, Low value).

Q&A visual: lets users type a natural-language question and returns a visual answer from the model.

• These visuals depend on a well-modeled, well-named semantic model to give good answers
• Key influencers can analyze both categorical outcomes and continuous measures
• Q&A quality improves with synonyms and a clean model

Microsoft Learn resource: The Key influencers visual in Power BI (opens in a new tab)

The Analytics pane adds analytical lines to a visual, and forecasting projects a trend forward, all without writing DAX.

What you need to know

• The Analytics pane adds constant, average, min, max, median, percentile, and trend lines
• Reference lines mark targets or thresholds; error bars show a range of uncertainty around a value
• Forecasting projects a line chart into the future with a confidence interval, on a time axis
• Availability of each analytics option depends on the visual type

Microsoft Learn resource: The Analytics pane in Power BI Desktop (opens in a new tab)

Anomaly detection finds unexpected values in a time series and can explain the likely cause, helping you spot problems in the data automatically.

What you need to know

• Anomaly detection is enabled in the Analytics pane on a line chart with a date axis
• It highlights points that fall outside an expected range and shades the expected boundary
• Selecting an anomaly shows possible explanations ranked by strength
• Sensitivity is adjustable to catch more or fewer anomalies

Microsoft Learn resource: Anomaly detection in Power BI (opens in a new tab)

Copilot can describe and summarize a semantic model in plain language, helping you and consumers understand what data is available and what it means.

What you need to know

• Copilot can summarize the model's tables, fields, and relationships in natural language
• It can help document a model and answer questions about what a field or measure represents
• Good table, column, and measure names and descriptions make the summary far more accurate
• Copilot features in Power BI require a supported Fabric or Premium capacity

Microsoft Learn resource: Copilot in Power BI overview (opens in a new tab)

Workspaces are the containers in the Power BI service where you collaborate on reports, semantic models, dashboards and other items before you distribute them. Understanding workspaces is the foundation of the whole service.

What you need to know

• A workspace holds related items: semantic models, reports, dashboards, dataflows and more
• My workspace is personal; shared workspaces are for team collaboration and app publishing
• Workspace settings cover the name, description, contacts, and the OneDrive/SharePoint location
• Publishing an app is done from a workspace, so a workspace is the unit you build and then share

Microsoft Learn resource: Create workspaces in Power BI (opens in a new tab)

An app is a polished, read-only package of selected content published from a workspace to a broad audience. It is the recommended way to distribute to many consumers.

What you need to know

• An app publishes chosen reports and dashboards from a workspace to consumers as one bundle
• Audiences let one app show different content to different groups of users
• Updating the workspace does not change the live app until you press Update app
• Consumers get the app; they do not see the workspace or its work-in-progress

Microsoft Learn resource: Publish an app in Power BI (opens in a new tab)

Exam tip: Changes made in a workspace are not visible to app users until you republish with Update app; expect a question that hinges on this.

Getting content into a workspace is usually a publish from Power BI Desktop, but you can also upload files directly. Republishing updates the item in place.

What you need to know

• Publish from Power BI Desktop sends the report and its semantic model to a chosen workspace
• Get data > Files can upload a .pbix, Excel workbook, or CSV directly in the service
• Republishing a report with the same name overwrites the existing item in the workspace
• A published .pbix creates two items: the report and its semantic model

Microsoft Learn resource: Publish reports from Power BI Desktop (opens in a new tab)

A dashboard is a single-page canvas in the service built from pinned tiles, often drawing on several reports and models. It differs from a report in important ways the exam checks.

What you need to know

• You build a dashboard by pinning visuals (tiles) from reports or Q&A
• Dashboards support data alerts on tiles; reports do not
• Dashboards are a service-only feature

Microsoft Learn resource: Create a dashboard in Power BI (opens in a new tab)

Power BI offers several ways to share, and picking the right one for the audience and governance need is a common scenario question.

What you need to know

• Apps: best for distributing to a large audience with a curated, read-only experience
• Share a report or dashboard: quick sharing with specific people via a link
• Workspace access: for collaborators who build content, through workspace roles
• Publish to web: creates a public, anonymous link, so never for confidential data
• Embedding and export cover apps and documents outside Power BI

Microsoft Learn resource: Power BI implementation planning: Content distribution and sharing (opens in a new tab)

Exam tip: Publish to web is fully public and anonymous; it must never be used for sensitive or internal data.

Subscriptions email a snapshot on a schedule, and data alerts notify you when a value crosses a threshold. Both keep people informed without opening Power BI.

What you need to know

• A subscription emails a report page or dashboard image on a set schedule
• Data alerts are set on dashboard tiles (card, KPI, gauge) and fire when a value crosses a threshold
• Alerts can notify in the service and by email, and can feed Power Automate flows
• Data alerts are a dashboard feature, not a report feature

Microsoft Learn resource: Set data alerts in the Power BI service (opens in a new tab)

Endorsement signals which content is trustworthy. The two levels, Promoted and Certified, help users find the right semantic model or report.

What you need to know

• Promoted is a lighter endorsement any contributor can apply to their content
• Certified is a stronger, controlled endorsement, granted only by users an admin authorizes
• Endorsed content is easier to find and marked with a badge in the data hub and lists
• Endorsement encourages reuse of governed semantic models over ad-hoc copies

Microsoft Learn resource: Endorse Power BI content (opens in a new tab)

A data gateway bridges the cloud service and data that stays on-premises or in a private network. Knowing when one is needed is frequently tested.

What you need to know

• A gateway is required to refresh or query data sources that are on-premises or behind a firewall
• Cloud sources (SharePoint Online, Azure SQL, and similar) generally do not need a gateway
• The on-premises data gateway (standard mode) supports scheduled refresh, DirectQuery, and live connection for multiple users
• A personal mode gateway supports Import refresh for one user only
• The gateway must map its data sources to the ones your semantic model uses

Microsoft Learn resource: What is an on-premises data gateway? (opens in a new tab)

Exam tip: If a scenario refreshes from an on-premises SQL Server or a local file share, it needs an on-premises data gateway; a purely cloud source does not.

Import models hold a cached copy of the data, so they must be refreshed to stay current. Scheduled refresh automates that in the service.

What you need to know

• Scheduled refresh runs on a set frequency and times; it applies to Import (and composite) models
• Pro workspaces allow up to 8 refreshes a day; Premium/Fabric capacity allows up to 48
• On-premises or gateway-backed sources need valid gateway connections and credentials
• DirectQuery and Direct Lake do not need a data refresh, since data is read at query time
• Refresh failure notifications can be sent to the model owner

Microsoft Learn resource: Configure scheduled refresh in Power BI (opens in a new tab)

Workspace roles decide what a collaborator can do inside a workspace. There are four, and the exam expects you to know what each one allows.

The four workspace roles:

1. Viewer: can view and interact with content, but not edit it
2. Contributor: can create and edit content in the workspace, but not publish or manage the app
3. Member: can do everything a Contributor can, plus publish/update the app and add others (up to Member)
4. Admin: full control, including workspace settings, deletion, and adding admins

• Roles are assigned to users or, better, to security groups
• Assigning to a group scales access management and reduces per-user churn
• Roles apply to the whole workspace, not to individual items

Microsoft Learn resource: Roles in workspaces in Power BI (opens in a new tab)

Sometimes you need to share a single report or grant access to one item rather than the whole workspace. Item-level sharing handles those targeted cases.

What you need to know

• Sharing a report or dashboard grants access to that item without workspace membership
• Share links can be for people in your org, specific people, or (if allowed) anyone with the link
• Share permissions can allow reshare and build on top of the underlying model
• Item-level access complements, but does not replace, workspace roles

Microsoft Learn resource: Share Power BI reports and dashboards (opens in a new tab)

Reusing one governed semantic model across many reports needs its own permissions. Build permission lets others create content on a shared model.

What you need to know

• Read permission lets a user consume reports on the model; Build lets them create new content from it
• Reshare and Build are granted through the model's Manage permissions page
• The data hub and Get data let users find and connect to models they can build on
• Granting Build on a certified model is how you promote a single version of the truth

Microsoft Learn resource: Manage semantic model access permissions (opens in a new tab)

Row-level security (RLS) restricts which rows of data a user can see, so one report shows each person only their own data. You define roles with DAX filters.

What you need to know

• RLS roles are created in Power BI Desktop under Manage roles, each with a DAX filter on a table
• A static role hardcodes the filter (for example, Region = "West")
• A dynamic role uses USERNAME() or USERPRINCIPALNAME() to filter by the signed-in user
• Test roles in Desktop with View as, and the filter flows to related tables through relationships
• RLS applies to Viewers; workspace Admins, Members and Contributors bypass it by default

Microsoft Learn resource: Row-level security (RLS) with Power BI (opens in a new tab)

Exam tip: Dynamic RLS keys off USERPRINCIPALNAME(), usually matched against a table that maps each user to the data they may see.

Defining an RLS role is only half the job; users or groups must be added to it in the service for the restriction to take effect.

What you need to know

• After publishing, add members to each role under Security on the semantic model in the service
• Assign security groups rather than individuals so membership scales
• A user not in any RLS role, with only view access, sees no data on a model that has RLS
• Members of the workspace with edit rights are not restricted by RLS

Microsoft Learn resource: Row-level security (RLS) with Power BI (opens in a new tab)

Sensitivity labels from Microsoft Purview classify and protect content, and they follow the data even when it is exported. They are how Power BI ties into organizational data protection.

What you need to know

• Labels (such as Confidential or Highly Confidential) come from Microsoft Purview Information Protection
• A label can enforce encryption and access restrictions that persist on export to Excel, PowerPoint, and PDF
• Labels can be applied to semantic models, reports, dashboards, and dataflows
• Labels can be inherited from the data source and can be mandated by policy
• Applying a label requires the feature to be enabled and the user to be licensed for it

Microsoft Learn resource: Sensitivity labels in Power BI (opens in a new tab)