Power BI Advanced Series · Module 1 — Lesson 2 · by Raushan Ranjan, MCT

Why Multi-Source Data Is the Real Enterprise Problem

Your CEO wants a single dashboard. Your sales data is in SQL Server. Inventory targets live in a SharePoint Excel file. Customer churn scores are in Azure SQL. Pipeline data is in Dynamics 365. No single source has the full picture — and manually copying between them every morning isn't a strategy, it's a time bomb.

This is the problem Power BI's data connectivity system was designed to solve. In this lesson you'll learn exactly how to pull from each source type, which connectivity mode to pick, and the four mistakes that silently break enterprise reports.

Quick Answer Power BI connects to 100+ source types across files, databases, cloud services, and online platforms. You choose a connectivity mode — Import, DirectQuery, or Live Connection — depending on data size, refresh needs, and governance requirements.
  • Import mode is fastest for reports but requires scheduled refresh
  • DirectQuery sends live queries — performance depends on source speed
  • Live Connection reuses a published Power BI dataset (no duplicate modelling)
  • Power Query is the transformation layer before data enters the model

The Universal Translator Analogy

Think of Power BI as a universal translator for business data. Just as a UN interpreter converts French, Mandarin, and Arabic into English without changing what was said, Power Query converts SQL rows, Excel ranges, JSON payloads, and OData feeds into a single unified table format — without altering the source systems. Your SQL Server doesn't know Power BI spoke to it. Your SharePoint file stays exactly as Finance left it. Power BI just listens, translates, and combines.

🏢 Enterprise Example — 500-Store Retail Chain

A national retailer with 500 stores needs a weekly executive dashboard. Their data landscape:

SQL Server (On-Prem)
POS transactions — 3M rows/day. Import mode with overnight refresh.
SharePoint Excel
Regional sales targets — updated monthly by Finance. Import mode.
Azure SQL Database
Customer loyalty scores — 2M customers. DirectQuery for near-real-time.
Dynamics 365
Open supplier orders — live operational data. DirectQuery.

Power BI connects to all four simultaneously. The analyst sees one model. Finance sees their targets next to actuals. Operations sees supplier delays against inventory. One dashboard. Four systems. Zero manual merging.

Source Categories: What Connects to What

CategoryExamplesWhen to Use
FilesExcel, CSV, JSON, PDF, XMLAd-hoc data from analysts; one-time imports
DatabasesSQL Server, PostgreSQL, Oracle, MySQL, AccessCore operational data; scheduled refresh or DirectQuery
Azure / CloudAzure SQL, Blob Storage, Synapse, DatabricksLarge-scale cloud workloads; modern data platform
Online ServicesSharePoint, Dynamics 365, Salesforce, Google AnalyticsSaaS platforms with OAuth connectors
Power PlatformPower BI Datasets, Dataflows, DataverseReuse certified models; avoid duplicate DAX
OtherWeb, OData, ODBC, OLEDB, Blank QueryAPIs, legacy systems, custom M queries
Pro tip: Always prefer a native connector (e.g., "SQL Server") over ODBC — native connectors use query folding, which means transformations run at the source rather than in Power BI memory.

Connectivity Modes: Import vs DirectQuery vs Live Connection

ModeData LocationRefreshBest ForLimitation
Import Copied into .pbix / dataset Scheduled (up to 8×/day free, 48×/day Premium) Fast visuals, complex DAX, small-to-medium data Data is never newer than last refresh
DirectQuery Stays at source Live — each visual interaction queries source Large tables, near-real-time, governance requirement Limited DAX functions; slow if source is slow
Live Connection Published Power BI / SSAS model Inherits from source dataset Multiple reports reusing one governed model No local transformations; model is read-only

For a deeper dive, see the dedicated article: Power BI Data Connectivity Modes Explained →

⚠️ 4 Mistakes That Break Multi-Source Reports
  1. Hardcoded file paths in Excel/CSV connectors. C:UsersRaushanDesktopsales.xlsx works on your machine, breaks on every other machine and in the cloud service. Fix: use a parameter or relative path, or move the file to SharePoint.
  2. Using DirectQuery for small, static tables. If your product dimension has 500 rows and never changes intraday, Import is always better. DirectQuery on small tables adds latency with zero benefit.
  3. Skipping scheduled refresh after publishing. You connect, publish, share the link — and forget to configure refresh. Three days later your CEO is looking at last week's data. Always configure the gateway and refresh schedule before sharing.
  4. Using the Web connector for REST APIs. The Web connector works for simple URLs, but for APIs that need auth headers, pagination, or POST requests, write a proper M function in Power Query or use a custom connector. The Web connector will quietly truncate or fail on page 2.
✅ Key Takeaways
  • Power BI connects to 100+ sources — categorised as files, databases, cloud, online services, and Power Platform
  • Import mode is fastest but has data freshness lag; DirectQuery is live but source-speed dependent
  • Live Connection lets multiple reports share one published dataset — the right choice for enterprise governance
  • Always prefer native connectors over ODBC to enable query folding and better performance
  • Avoid hardcoded file paths and always configure refresh before publishing to colleagues
🧠 Check Your Understanding

Q1. Your source table has 50 million rows and your business needs data no more than 15 minutes old. Which connectivity mode should you choose?

Show Answer

DirectQuery. Import mode can refresh at most 48×/day on Premium (every 30 min). For 15-minute freshness with 50M rows, DirectQuery is the correct choice — as long as your source (e.g., Azure SQL, Synapse) can handle the query load.

Q2. Three different report authors each build separate .pbix files connecting to the same SQL Server sales table. What is the problem with this approach, and what is the better solution?

Show Answer

Problem: Each author defines their own measures, potentially with inconsistent logic (e.g., different revenue calculations). Three datasets duplicate storage and refresh load.
Solution: One author builds and publishes the certified dataset. The other two use Live Connection to that dataset, so all three reports share identical logic and a single refresh.

Q3. A colleague shares a .pbix file with a CSV connector. When you open it you get a "File not found" error. What is the most likely cause?

Show Answer

The connector uses a hardcoded local file path (e.g., C:UsersJohnDocumentsdata.csv) that exists only on their machine. Fix: move the file to a shared location (SharePoint, OneDrive, network path) and update the source path, or use a Power Query parameter so it can be changed without editing the query.