Q1: What are Virtual Tables in DAX and how are they used?
Answer: Virtual Tables are in-memory tables created dynamically during DAX evaluation — they do not physically exist in the data model. Functions like FILTER, ADDCOLUMNS, SUMMARIZE, SELECTCOLUMNS, UNION, INTERSECT, EXCEPT, DATATABLE, and GENERATEALL return virtual tables. They exist only during query execution and are discarded afterward. Virtual tables are commonly used inside CALCULATE as filter arguments, inside SUMX/AVERAGEX as the table argument for custom iterations, and for creating intermediate calculation results within complex measures.
🎯 Explain: Virtual Tables woh tables hain jo DAX ke andar temporarily banti hain — model mein permanently nahi hoti. FILTER(Sales, Sales[Region]="North") ek virtual table return karta hai — sirf North ki rows. Ye table memory mein exist karti hai sirf us query ke dauran. Use cases: CALCULATE ke andar complex filters, SUMX ke andar filtered table iterate karna, SUMMARIZE se grouped data banana. Advanced DAX likhne ke liye Virtual Tables samajhna zaroori hai — yeh building blocks hain.

Q2: What is CALCULATETABLE and how is it different from CALCULATE?
Answer: CALCULATE evaluates a scalar expression (returns a single value) in a modified filter context. CALCULATETABLE evaluates a table expression (returns a table) in a modified filter context. CALCULATETABLE is used when you need to apply filters to a table-returning function. For example, CALCULATETABLE(VALUES(Products[Category]), Sales[Region]="North") returns distinct product categories sold in North region only. CALCULATETABLE is essential for creating filtered virtual tables that feed into other DAX calculations.
🎯 Explain: CALCULATE = ek single value return karta hai modified filters ke saath — jaise filtered SUM. CALCULATETABLE = ek table return karta hai modified filters ke saath — jaise filtered list of products. Example: CALCULATE(SUM(Sales[Amount]), Region="North") → ek number. CALCULATETABLE(VALUES(Products[Name]), Region="North") → ek table of product names. CALCULATETABLE SUMMARIZE, VALUES, FILTER ke saath use hota hai jab filtered table chahiye. Advanced measures mein CALCULATETABLE frequently use hota hai.

Q3: What is TREATAS in DAX and when should you use it?
Answer: TREATAS applies the data lineage of target columns to a virtual table — essentially treating one column's values as if they belong to another column for filtering purposes. It creates a virtual relationship without needing a physical model relationship. Syntax: CALCULATE(SUM(Sales[Amount]), TREATAS(VALUES(Budget[ProductID]), Sales[ProductID])). Use cases: filtering across unrelated tables, applying budget filters to actuals when both tables have no direct relationship, and creating many-to-many bridges dynamically without bridge tables.
🎯 Explain: TREATAS ek powerful function hai jo bina physical relationship ke ek table ke values ko dusri table ke column pe filter ki tarah lagata hai. Maan lo Budget table aur Sales table hain — dono mein ProductID hai lekin relationship nahi hai. TREATAS(VALUES(Budget[ProductID]), Sales[ProductID]) — Budget ke products ko Sales pe filter laga dega. Virtual relationship jaisa kaam karta hai. Real-world mein Budget vs Actual comparison, unrelated table filtering, aur complex multi-fact table scenarios mein use hota hai.

Q4: What is the difference between VALUES, DISTINCT, and ALL in DAX?
Answer: VALUES returns distinct values from a column respecting the current filter context and includes BLANK if present due to referential integrity violations. DISTINCT returns distinct values respecting filters but excludes the extra BLANK row. ALL returns all distinct values from a column ignoring any active filters — it removes filter context. Key differences: VALUES and DISTINCT respect filters (VALUES includes BLANK, DISTINCT excludes it), ALL ignores all filters. VALUES is used for dynamic filtering, ALL for percentage-of-total calculations where you need the grand total regardless of filters.
🎯 Explain: VALUES = unique values current filter ke andar + BLANK row agar referential integrity break ho. DISTINCT = unique values current filter ke andar — BLANK row nahi. ALL = sab unique values — filters ignore — jaise poora Products list chahe slicer mein kuch bhi ho. Example: Slicer mein "North" hai → VALUES(Sales[Product]) sirf North ke products dega. ALL(Sales[Product]) sab products dega chahe filter kuch bhi ho. % of Total calculation mein: DIVIDE(SUM(Sales[Amount]), CALCULATE(SUM(Sales[Amount]), ALL(Sales))) — denominator mein ALL use hoga.

Q5: How does RANKX work and what are its parameters?
Answer: RANKX ranks each row in a table based on an expression. Syntax: RANKX(table, expression, [value], [order], [ties]). Parameters: (1) table — the table to iterate and rank (usually ALL(DimTable)). (2) expression — the value to rank by (usually a measure). (3) value — optional, the specific value to rank (defaults to current context). (4) order — ASC or DESC (default DESC — highest rank 1). (5) ties — DENSE (no gaps after ties) or SKIP (gaps after ties, default). Common pattern: RANKX(ALL(Products), [Total Sales],, DESC, DENSE).
🎯 Explain: RANKX products ya employees ko rank karta hai based on sales, performance, etc. RANKX(ALL(Products[Name]), [Total Sales]) — har product ko uski Total Sales ke basis pe rank dega — sabse zyada sales = Rank 1. ALL use karte hain taaki ranking sab products pe ho, sirf filtered pe nahi. DESC = highest first. DENSE ties mein gaps nahi chhodta — 1,2,2,3 (DENSE) vs 1,2,2,4 (SKIP). Dashboard mein "Top 10 Products" dikhane ke liye RANKX use hota hai. VAR ke saath combine karo readable measure ke liye.

Q6: What is SUMMARIZE vs SUMMARIZECOLUMNS and which should you use?
Answer: Both group data by specified columns, but they differ significantly. SUMMARIZE creates a grouped summary table within the current filter context — it can add calculated columns using ADDCOLUMNS. SUMMARIZECOLUMNS is optimized by the engine, handles BLANK rows better, automatically removes rows where all measures are BLANK, and accepts filter parameters directly. SUMMARIZECOLUMNS is generally faster and preferred for measure calculations. However, SUMMARIZECOLUMNS cannot be used inside ADDCOLUMNS or other row context scenarios. Use SUMMARIZE for Calculated Tables, SUMMARIZECOLUMNS for internal query generation (Power BI uses it automatically).
🎯 Explain: SUMMARIZE = GROUP BY jaisa — SUMMARIZE(Sales, Sales[Region], "Total", SUM(Sales[Amount])) — region wise total. SUMMARIZECOLUMNS = engine-optimized version — faster, BLANK rows automatically remove karta hai. SUMMARIZECOLUMNS Power BI internally visuals ke liye use karta hai. Rule: Calculated Table banani hai → SUMMARIZE use karo. Measure ke andar complex grouping → ADDCOLUMNS + SUMMARIZE pattern. Generally SUMMARIZECOLUMNS avoid karo direct measures mein — limitations hain nested contexts mein.

Q7: How does Filter Propagation work in Power BI data models?
Answer: Filter Propagation is the mechanism by which filters applied on one table automatically flow to related tables through relationships. In a Star Schema, filters propagate from Dimension tables (one side) to Fact tables (many side) — this is the default single-direction flow. The direction is determined by the cross-filter direction setting. With bi-directional (Both) filtering, filters can also propagate from Fact to Dimension — but this is not recommended unless specifically needed as it can cause ambiguity and performance issues in complex models.
🎯 Explain: Filter Propagation matlab ek table pe filter lagao toh related table automatically filter ho jaaye. Products table mein "Laptop" select kiya → Sales table mein sirf Laptop ki rows dikhegi. Yeh Dimension → Fact direction mein hota hai (one-to-many ki "one" side se "many" side). Both direction mein ulta bhi chalega — lekin complex models mein problems aate hain — circular dependency, ambiguous paths. Best practice: Single direction rakho, zaroorat ho toh specific measures mein CROSSFILTER() use karo.

Q8: What is the CALCULATE filter evaluation order?
Answer: CALCULATE evaluates filters in a specific order: (1) First, the outer filter context (from slicers, visuals, page filters) is established. (2) Then, CALCULATE's filter arguments are evaluated — they can override, add to, or modify the existing context. (3) If Context Transition occurs (measure inside iterator), row context values become filters. (4) ALL/REMOVEFILTERS remove existing filters before new ones are applied. (5) The expression is then evaluated in this final modified context. Understanding this order is crucial for debugging unexpected DAX results.
🎯 Explain: CALCULATE ka evaluation order samajhna DAX debugging ke liye critical hai. Step 1: Bahar ke filters set hote hain (slicer, visual). Step 2: CALCULATE ke andar jo filters likhe hain woh apply hote hain — yeh outer filters ko override kar sakte hain. Step 3: Context Transition agar ho toh row values filters ban jaate hain. Step 4: ALL pehle clear karta hai, phir naye filters lagte hain. Step 5: Final context mein expression evaluate hota hai. Galat order samajhoge toh measures galat results denge — yeh most common advanced DAX debugging issue hai.

Q9: What is the difference between SELECTEDVALUE, HASONEVALUE, and ISFILTERED?
Answer: SELECTEDVALUE returns the single value in a column if only one distinct value exists in the current filter context, otherwise returns an alternate result. HASONEVALUE returns TRUE/FALSE indicating whether exactly one distinct value exists in the filter context. ISFILTERED returns TRUE if a column has any direct filter applied (even if multiple values are selected). Use SELECTEDVALUE for dynamic measure titles and conditional logic. HASONEVALUE for conditional branching in measures. ISFILTERED to detect if a slicer or filter is actively filtering a column.
🎯 Explain: SELECTEDVALUE = agar slicer mein sirf ek value selected hai toh woh return karo, warna alternate value. Example: SELECTEDVALUE(Products[Name], "All Products") — ek product select hai toh naam aayega, multiple hain toh "All Products". HASONEVALUE = TRUE/FALSE — conditional logic ke liye. IF(HASONEVALUE(Products[Category]), "Single", "Multiple"). ISFILTERED = kya koi filter laga hai column pe — chahe ek value ho ya multiple. Dynamic titles aur conditional measures mein bahut use hota hai.

Q10: What is CROSSFILTER function in DAX?
Answer: CROSSFILTER modifies the cross-filter direction of a relationship within a specific CALCULATE expression — without permanently changing the model relationship. Syntax: CALCULATE(expression, CROSSFILTER(Table1[Column], Table2[Column], direction)). Direction can be NONE (disable filtering), ONEWAY (single direction), ONEWAY_RIGHTFILTERSLEFT, ONEWAY_LEFTFILTERSRIGHT, or BOTH (bi-directional). This is useful when you need bi-directional filtering for a specific measure without setting the entire relationship to Both in the model.
🎯 Explain: Model mein relationship Single direction hai — lekin ek specific measure mein Both direction chahiye. Model change karna risky hai — toh CROSSFILTER use karo. CALCULATE(COUNTROWS(Products), CROSSFILTER(Sales[ProductID], Products[ProductID], BOTH)) — sirf is measure ke liye Both direction activate hoga. Baaki sab measures mein Single rahega. NONE use karo jab temporarily relationship disable karni ho. Very powerful aur safe — model untouched rehta hai.

Q11: How do you optimize slow DAX measures?
Answer: DAX optimization strategies include: (1) Use VAR to avoid redundant calculations — store results once, reference multiple times. (2) Replace iterators on large tables with simpler aggregations where possible. (3) Use KEEPFILTERS instead of FILTER inside CALCULATE — KEEPFILTERS intersects with existing context rather than overriding. (4) Avoid DISTINCTCOUNT on high-cardinality columns — use approximate counting if exact count is not needed. (5) Use DAX Studio with Server Timings to identify Storage Engine (SE) vs Formula Engine (FE) bottlenecks. (6) Pre-aggregate data in Power Query instead of complex DAX calculations. (7) Avoid complex nested CALCULATE patterns.
🎯 Explain: Slow DAX ka root cause dhundho pehle — DAX Studio mein query run karo, Server Timings dekho. Storage Engine (SE) slow hai toh data model optimize karo — columns kam karo, cardinality reduce karo. Formula Engine (FE) slow hai toh DAX formula simplify karo. VAR se repeated calculations avoid karo. SUMX 50 lakh rows pe slow hoga — agar possible hai toh pre-aggregate karo Power Query mein. KEEPFILTERS FILTER se faster hai CALCULATE ke andar. Step-by-step approach follow karo — blindly optimize mat karo.

Q12: What is KEEPFILTERS and how is it different from FILTER inside CALCULATE?
Answer: When you use a direct column filter inside CALCULATE like Sales[Region]="North", it OVERRIDES any existing filter on that column. KEEPFILTERS wraps around the filter to make it INTERSECT with existing filters instead of overriding. Example: CALCULATE(SUM(Sales[Amount]), KEEPFILTERS(Sales[Region]="North")). If a slicer has "North" and "South" selected, direct filter would show only North (override). KEEPFILTERS would show North only if North is in the slicer selection — if slicer has only "South", result would be BLANK. KEEPFILTERS is safer and more predictable in complex filter scenarios.
🎯 Explain: Direct filter CALCULATE mein = override — purana filter hata ke naya lagao. KEEPFILTERS = intersection — purana aur naya dono match hona chahiye. Practical example: Slicer mein "South" selected hai. CALCULATE(SUM(...), Region="North") → North ki sales dikhayega (override). CALCULATE(SUM(...), KEEPFILTERS(Region="North")) → BLANK dikhayega kyunki slicer mein South hai aur KEEPFILTERS North maang raha hai — dono match nahi karte. KEEPFILTERS predictable behavior deta hai — advanced measures mein recommended hai.

Q13: What are Composite Models in Power BI?
Answer: Composite Models allow a single Power BI report to combine data from multiple storage modes — Import, DirectQuery, and Dual — within the same data model. This means you can have some tables imported into memory (fast, cached) and other tables connected via DirectQuery (real-time, live). Dual mode tables can act as either Import or DirectQuery depending on the query context. Composite Models are ideal when you need a mix of performance (Import for dimensions) and real-time data (DirectQuery for large fact tables).
🎯 Explain: Pehle ya toh poora Import hota tha ya poora DirectQuery — ab dono mix kar sakte ho. Example: Products table (5000 rows) → Import mode (fast). Sales table (5 crore rows) → DirectQuery (real-time). Dono ek hi report mein. Dual mode tables dono jaise kaam kar sakti hain depending on context. Real-world mein bahut useful hai — large enterprise data mein Import se memory issue aata hai, DirectQuery se speed issue — Composite Model dono ka best deta hai.

Q14: What are Power BI Dataflows?
Answer: Dataflows are a cloud-based data preparation technology in Power BI Service that allows you to define ETL (Extract, Transform, Load) processes using Power Query Online. Data is extracted from sources, transformed using familiar Power Query steps, and stored in Azure Data Lake Storage (or Common Data Model folders). Multiple reports and datasets can then connect to the same Dataflow — ensuring data consistency and reducing redundant data preparation. Dataflows promote a single source of truth and reusable data pipelines across the organization.
🎯 Explain: Normally har report mein separately data connect, clean, load karte ho — same kaam repeatedly. Dataflows mein ek baar cloud mein data prepare karo — phir multiple reports us ek Dataflow se data le lein. Ek centralized data preparation layer. Example: HR data Dataflow mein clean karo — HR Report, Attendance Report, Payroll Report — sab same clean data use karenge. Data consistency maintain hoti hai. IT team Dataflows manage karti hai, report creators sirf connect karke reports banate hain.

Q15: What are Deployment Pipelines in Power BI?
Answer: Deployment Pipelines provide a structured workflow for managing the lifecycle of Power BI content across three stages: Development, Test, and Production. Content is created in the Development stage, validated in Test, and promoted to Production for end users. This follows standard software development practices (Dev → QA → Prod). Changes are deployed between stages with a single click, and comparison views show differences between stages. Deployment Pipelines require Power BI Premium capacity and help prevent untested reports from reaching production.
🎯 Explain: Jaise software development mein Dev → Testing → Production hota hai — Power BI mein bhi same. Development mein report banao aur changes karo. Test mein QA team verify kare. Production mein end users dekhein. Bina Deployment Pipelines ke log directly production mein changes karte the — risky tha. Ab structured process hai — ek click mein Dev se Test, Test se Prod promote karo. Premium feature hai. Enterprise mein mandatory practice hai — interview mein mention karo toh corporate experience dikhta hai.

Q16: What is XMLA Endpoint in Power BI Premium?
Answer: XMLA (XML for Analysis) Endpoint is a protocol that allows external tools to connect to Power BI Premium datasets as if they were Analysis Services databases. With XMLA read/write enabled, you can use third-party tools like Tabular Editor, DAX Studio, SQL Server Management Studio (SSMS), and ALM Toolkit to manage, develop, and query Power BI datasets. XMLA enables advanced scenarios like automated deployment, dataset documentation, programmatic measure management, and advanced debugging that are not possible through the Power BI Desktop interface alone.
🎯 Explain: XMLA Endpoint Power BI Premium datasets ko bahar ke tools se access karne deta hai. DAX Studio se directly dataset query karo debugging ke liye. Tabular Editor se measures bulk mein add/edit karo — GUI se ek ek karke karne se zyada fast. SSMS se dataset manage karo jaise SQL database manage karte ho. ALM Toolkit se deployments automate karo. Yeh enterprise-level development ke liye game-changer hai — interview mein "I use Tabular Editor via XMLA for efficient measure management" — senior-level answer hai.

Q17: What is the difference between Power BI Pro, Premium Per User (PPU), and Premium Per Capacity?
Answer: Power BI Pro is a per-user license that enables sharing and collaboration — each user who views shared content needs Pro. PPU (Premium Per User) provides Premium features at a per-user cost — includes larger models, Deployment Pipelines, Dataflows, XMLA endpoints — but all viewers also need PPU. Premium Per Capacity is an organizational subscription providing dedicated cloud resources — Pro users can share content with free users, no per-viewer license needed. Premium Per Capacity is best for wide distribution, PPU for small teams needing premium features, Pro for standard collaboration.
🎯 Explain: Pro = basic sharing — har viewer ko Pro license chahiye — monthly per user cost. PPU = Premium features per user — Deployment Pipelines, larger datasets — but viewers ko bhi PPU chahiye — chhoti team ke liye sahi. Premium Capacity = organization-level — dedicated resources — Pro users free users ko bhi share kar sakte hain — badi company ke liye. Simple rule: 50 log dekhenge → Pro (sab ko Pro do). 5 log premium features chahiye → PPU. 500+ log dekhenge → Premium Capacity (free users bhi dekh sakte hain).

Q18: What are Datamart in Power BI?
Answer: Datamarts are self-service relational databases in Power BI that allow business users to create and manage their own data repositories without IT dependency. They provide a fully managed Azure SQL DB behind the scenes, support SQL querying, and allow visual query design using a no-code interface. Datamarts combine the power of a relational database with the ease of Power BI — users can define tables, relationships, write T-SQL queries, and connect datasets to the Datamart. They bridge the gap between self-service analytics and proper data management.
🎯 Explain: Datamarts self-service mini-databases hain Power BI ke andar. Business users apna data dal sakte hain, tables bana sakte hain, SQL likh sakte hain — bina IT team ke help ke. Backend mein Azure SQL DB chalta hai — fully managed. Power BI ke visual interface se query design kar sakte ho — no-code. Reports Datamart se directly connect hote hain. Use case: Finance team apna budget data Datamart mein rakh le, SQL se query kare, reports banaye — IT pe dependent na ho. Self-service analytics ka next level hai.

Q19: What are Paginated Reports in Power BI?
Answer: Paginated Reports are pixel-perfect, print-ready reports designed for precise formatting and multi-page output. Unlike interactive Power BI reports that are optimized for screen exploration, Paginated Reports are optimized for printing and PDF export — they can span hundreds of pages with exact header/footer placement, page breaks, and detailed table layouts. They are created using Power BI Report Builder (based on SQL Server Reporting Services) and published to Power BI Service. They require Premium capacity and are ideal for invoices, financial statements, and regulatory reports.
🎯 Explain: Normal Power BI reports interactive hain — visuals click karo, filter karo. Paginated Reports print ke liye hain — fixed layout, page numbers, headers/footers, ek invoice 10 pages ka bhi ho sakta hai. HR payslips, bank statements, inventory lists — jahan exact formatting chahiye wahan Paginated Reports use hote hain. Power BI Report Builder se banate hain (alag tool hai). Premium feature hai. Interview mein "I use Paginated Reports for financial statements and invoices that need pixel-perfect formatting" — yeh use case batao.

Q20: What are AI Visuals in Power BI?
Answer: Power BI includes built-in AI-powered visuals: (1) Key Influencers — identifies factors that most influence a selected metric. (2) Decomposition Tree — allows interactive drill-down into contributing factors hierarchically. (3) Q&A Visual — enables natural language questions to generate visualizations automatically. (4) Smart Narratives — auto-generates text descriptions of data insights. (5) Anomaly Detection — highlights unexpected spikes or dips in time series data. These AI visuals democratize data analysis by enabling non-technical users to discover insights without writing DAX.
🎯 Explain: AI Visuals Power BI ke built-in smart tools hain. Key Influencers — "kya factors high churn cause kar rahe hain?" — automatically batayega (salary low, tenure short). Decomposition Tree — interactive drill-down — Total Sales kaise break hoti hai Region → Product → Quarter. Q&A — type karo "total sales by region" — chart ban jayega. Smart Narratives — data ka automatic text summary. Anomaly Detection — sudden spike ya drop highlight karega timeline mein. Non-technical managers ke liye perfect — bina DAX ke insights milte hain.

Q21: What are What-If Parameters in Power BI?
Answer: What-If Parameters create a table of numeric values with a slicer that allows users to dynamically change a variable and see its impact on calculations. For example, a "Discount %" parameter with range 0% to 50% lets users slide the value and instantly see how different discount levels affect profit. Power BI creates a calculated table, a measure for the selected value, and a slicer automatically. What-If Parameters enable scenario analysis, sensitivity testing, and interactive financial modeling within reports.
🎯 Explain: What-If Parameter ek slider deta hai users ko — value change karo aur calculations live update ho jaayein. Example: "Growth Rate %" parameter banao 5% se 25% tak — slider pe 15% select karo toh forecast measures automatically 15% growth factor se calculate honge. Modeling tab → New Parameter → range set karo → done. Power BI background mein ek table banata hai values ki aur ek measure jo selected value return karta hai. Financial modeling, sales forecasting, pricing analysis — bahut powerful feature hai. Interview mein scenario analysis ke context mein batao.

Q22: What is Field Parameters feature in Power BI?
Answer: Field Parameters allow users to dynamically switch which fields (columns or measures) are used in a visual through a slicer. For example, you can create a Field Parameter containing Sales Amount, Profit, and Quantity measures — users select which measure to display in a chart via slicer without needing multiple charts. Field Parameters create a Calculated Table with metadata about included fields and a slicer for selection. This enables dynamic axis switching, measure switching, and highly flexible report designs with fewer visuals.
🎯 Explain: Pehle agar user ko Sales Amount, Profit, Quantity alag alag dekhni hoti — teen charts banane padte. Field Parameters se ek chart banao — slicer mein "Sales Amount" ya "Profit" select karo — same chart dynamically change ho jayega. Modeling tab → New Parameter → Fields select karo. Slicer automatically ban jaata hai. Report clean rehta hai — kam visuals, zyada flexibility. Dynamic reports ke liye must-have feature hai. Interview mein "I use Field Parameters for dynamic measure switching to reduce visual count and improve user experience" — practical answer hai.

Q23: What is Calculation Groups in Power BI?
Answer: Calculation Groups allow you to define a set of calculation items that apply transformations to existing measures dynamically. Instead of creating separate measures for Current Year, Previous Year, YoY %, YTD for each base measure (Sales, Profit, Cost), you create one Calculation Group with items like "CY", "PY", "YoY%", "YTD" — and they automatically apply to any measure placed in a visual. Calculation Groups are created using Tabular Editor (via XMLA endpoint) and dramatically reduce the number of measures needed in complex models.
🎯 Explain: Problem: 5 base measures (Sales, Profit, Cost, Quantity, Orders) × 4 time calculations (CY, PY, YoY%, YTD) = 20 measures likhne padenge. Calculation Groups se: 5 base measures + 1 Calculation Group (4 items) = same result. Calculation Group ek reusable template jaisa hai — CY, PY, YoY% items define karo — woh automatically har measure pe apply ho jaayenge. Tabular Editor se banate hain (XMLA endpoint chahiye). Enterprise projects mein measures 100+ se 20-30 pe aa jaate hain. Yeh senior-level feature hai — interview mein mention karo toh expert impression padega.

Q24: How do you handle Slowly Changing Dimensions (SCD) in Power BI?
Answer: Slowly Changing Dimensions track historical changes in dimension data over time. Three types: SCD Type 1 — overwrite old values (no history). SCD Type 2 — add new rows with versioning (effective dates, current flags). SCD Type 3 — add new columns for old/new values. In Power BI: SCD Type 1 is handled naturally through refresh. SCD Type 2 requires the source to maintain history rows with surrogate keys, start/end dates — Power BI connects to these versioned rows and uses them for historical reporting. Dataflows with incremental refresh can also help manage SCD Type 2 scenarios.
🎯 Explain: SCD = dimension data slowly change hota hai. Employee ka department change ho gaya — purana record kya karo? Type 1: Simply overwrite — Department IT se HR ho gaya, bas update karo. History lost. Type 2: Nayi row add karo — Employee ka IT wala record bhi rahe aur HR wala bhi — StartDate, EndDate, IsCurrent flag ke saath. History preserved. Type 3: Original_Dept aur Current_Dept dono columns mein rakho — limited history. Power BI mein Type 2 sabse common hai — source mein history maintain honi chahiye. Power BI sirf consume karta hai — source pe SCD logic lagta hai.

Q25: How would you design a Power BI solution for a large enterprise with 10,000+ users?
Answer: Enterprise solution design includes: (1) Premium Capacity for wide distribution to free users. (2) Star Schema data model with Composite Models — Import for dimensions, DirectQuery for large facts. (3) Centralized Dataflows for shared data preparation. (4) Deployment Pipelines for Dev-Test-Prod workflow. (5) Dynamic RLS for user-level security. (6) Incremental Refresh for efficient data loading. (7) Aggregation Tables for dashboard performance. (8) Calculation Groups to reduce measure count. (9) XMLA + Tabular Editor for professional development. (10) Monitoring via Premium Capacity Metrics app and Usage Metrics reports.
🎯 Explain: 10,000 users ke liye architecture plan karna padta hai — sirf report banana kaafi nahi. Premium Capacity lena padega taaki free users bhi access kar sakein. Composite Models se large data handle karo. Dataflows se data centralize karo. Deployment Pipelines se safe deployments. RLS se security. Incremental Refresh se efficient loading. Aggregation Tables se fast dashboards. Monitoring dashboards se capacity usage track karo. Yeh complete enterprise BI architecture hai — interview mein yeh structured approach batao.

Q26: What is Power BI Embedded and how does it differ from Power BI Service?
Answer: Power BI Embedded is an Azure service that allows developers to embed interactive Power BI reports and dashboards into custom applications, portals, and websites using APIs. Key differences: Service is a standalone cloud platform for business users, Embedded is for developers integrating Power BI into their own apps. Embedded uses capacity-based pricing (Azure resource), Service uses per-user licensing. Embedded supports two scenarios: (1) Embed for your organization (users have Power BI licenses). (2) Embed for your customers (app owns the data, customers do not need Power BI licenses).
🎯 Explain: Power BI Service = standalone platform — users browser mein login karke reports dekhte hain. Power BI Embedded = apni app mein Power BI reports embed karo — user ko Power BI ka pata bhi nahi chalta. Example: SaaS application mein analytics dashboard dena hai customers ko — Embedded use karo. "Embed for customers" mein customers ko Power BI license nahi chahiye — app handle karti hai authentication. ISVs (Independent Software Vendors) ke liye perfect hai — apni app mein analytics add karo bina customers ko Power BI subscription dilaaye.

Q27: How do you handle data governance in Power BI?
Answer: Data governance in Power BI includes: (1) Sensitivity Labels — classify content as Public, General, Confidential using Microsoft Information Protection. (2) Data Loss Prevention (DLP) policies — prevent sharing of sensitive datasets. (3) Endorsement — certify or promote trusted datasets so users know which data sources are reliable. (4) Lineage View — visualize data flow from source to dashboard. (5) Impact Analysis — see which reports/dashboards are affected when a dataset changes. (6) Audit Logs — track user activities through Microsoft 365 compliance center. (7) Governance Policies — define who can publish, share, export data.
🎯 Explain: Governance matlab data ka control aur security. Sensitivity Labels se data classify karo — Confidential data pe extra restrictions lagao. Endorsement se official datasets mark karo — users trusted data use karein. Lineage View se dekho data source se report tak ka flow. Impact Analysis se pata chale ki dataset change karne se kaun kaun si reports affect hongi. Audit Logs se track karo kaun kya kar raha hai. Enterprise mein governance ke bina chaos hota hai — 100 log 100 datasets bana denge bina standards ke. Interview mein governance mention karna maturity dikhata hai.

Q28: What is the difference between Live Connection and DirectQuery?
Answer: Both avoid importing data into Power BI, but they differ in what source they connect to. DirectQuery connects to relational databases (SQL Server, Oracle) and sends DAX queries translated to SQL. You can add calculated columns, measures, and custom tables in the model. Live Connection connects to Analysis Services (SSAS Tabular/Multidimensional) or Power BI Service datasets — the model already exists on the server. With Live Connection, you cannot modify the data model in Desktop (no adding tables/columns) — you can only add report-level measures. Live Connection is "read-only model", DirectQuery is "build model on remote data".
🎯 Explain: DirectQuery = database se directly query karo — SQL Server, PostgreSQL. Power BI DAX ko SQL mein translate karta hai. Model Desktop mein modify kar sakte ho — columns, measures add karo. Live Connection = Analysis Services ya Power BI dataset se connect karo — model already bana hua hai server pe. Desktop mein model change nahi kar sakte — sirf report-level measures add kar sakte ho. DirectQuery = "I build the model, data stays remote." Live Connection = "Model already built elsewhere, I just consume it." Dono mein data import nahi hota.

Q29: How do you handle multi-currency conversion in Power BI?
Answer: Multi-currency conversion typically requires: (1) A Currency Exchange Rate table with Date, SourceCurrency, TargetCurrency, and ExchangeRate columns. (2) Relationships: Exchange Rate table relates to Date table and Sales table through currency and date columns. (3) DAX measure: Converted Amount = SUMX(Sales, Sales[Amount] * RELATED(ExchangeRates[Rate])). (4) For different reporting currencies, a What-If parameter or slicer lets users select the target currency. (5) Historical rates vs spot rates should be handled based on business requirements — use the rate valid on the transaction date for accurate conversion.
🎯 Explain: Multi-currency real-world mein bahut common hai — global companies mein. Step 1: Exchange Rate table banao — Date wise rates (USD to INR, EUR to INR). Step 2: Sales table se relate karo Date aur Currency ke through. Step 3: SUMX se har row pe conversion karo — Sales[Amount] * RELATED(ExchangeRates[Rate]). Step 4: Slicer do user ko target currency choose karne ke liye. Important: transaction date ke rate use karo (historical), aaj ka rate nahi (spot) — accounting standards ke liye zaroori. SUMX yahan perfect hai kyunki row-level conversion chahiye.

Q30: Describe a complex Power BI project you have worked on — challenges and solutions.
Answer: This is a behavioral/scenario question. Structure your answer using the STAR method — Situation, Task, Action, Result. Example: Situation — "Our company had 50+ Excel reports across 5 departments with inconsistent data." Task — "Consolidate everything into a unified Power BI analytics platform." Action — "I designed a Star Schema model, created centralized Dataflows for data preparation, implemented Dynamic RLS for department-level security, used Composite Models for a 100M+ row fact table, built Calculation Groups to manage 60+ time intelligence measures, and set up Deployment Pipelines for safe releases." Result — "Reduced reporting time from 3 days to real-time, achieved 98% user adoption, and saved 200+ man-hours monthly."
🎯 Explain: Yeh question har senior interview mein aata hai. STAR method use karo: (S) Company mein problem kya tha — scattered reports, inconsistent data. (T) Tumhara role kya tha — unified platform banana. (A) Kya kiya — Star Schema, Dataflows, RLS, Composite Models, Calculation Groups. (R) Result kya mila — time saved, adoption rate, man-hours reduced. Specific numbers do — "100M rows", "200 hours saved", "98% adoption." Bina numbers ke answer weak lagta hai. Even agar tumne itna bada project nahi kiya — apne experience ko structure karo STAR mein — professional dikhega.