Objective and core distinction
An analytics design links sources, ingestion, processing, analytical storage and reporting. ETL transforms before loading the target; ELT loads first and transforms afterward. A data lake can retain large varied file collections. A warehouse uses an analytical relational schema, and a lakehouse combines lake storage with table/query abstractions. These are roles and patterns, not guarantees that every implementation behaves identically. A summary result can answer a question while discarding detail needed for another task. Before choosing a design, write which original facts must remain, where transformations occur, and what output consumers actually require.
Worked example
Fictional case:a town archive receives image scans, JSON descriptions and fixed-column donation exports. Its future research questions are not all known. It retains originals in a file-oriented lake, then prepares a curated table for a monthly report. The original collection and curated table have different purposes. If processing happens before that target table is loaded, the relevant order is ETL; loading raw input first and processing in the target is ELT. Trace the actual sequence rather than choose from familiarity with the labels. A later total by donor category is an analytical measure over a selected grouping, not another copy of the original scan.
Changed condition and misconception
Microsoft Fabric integrates experiences such as Data Factory, engineering, warehousing, real-time work and reporting around OneLake, its logical data lake. Azure Databricks is another cloud analytics platform supporting large-scale engineering and data science with Delta Lake. Do not mistake a shared platform for one chart or one record. Likewise, resource integration does not establish the actual permissions, capacity, license or measured performance of a project. This foundation lesson identifies service purposes without starting resources. When a table supplies numeric measures, keep filtering, grouping and aggregation distinct:adding every row, counting rows and taking an average answer different questions.
Transfer check
Transfer check:a fictional team keeps only its final quarterly chart and discards the source files, then claims it can rebuild any future analysis. Which assumption should be challenged? Check basis:the chart may omit the original detail needed for a new question. Specify what source or intermediate facts must be retained to reproduce the intended work. Separately, read a pipeline described as loading then transforming and explain why the order matters. For an aggregation, write the selected group and operation before calculating. The formal tasks cover a pipeline order, a raw-file requirement, a platform foundation and a tiny supplied group total.
Try it, then explain it.
Complete a short set without looking at the example. After each answer, explain which evidence or rule supports your decision and why the alternatives do not fit.
Sources & authorship
Original OpenExam explanations and examples. References support the concepts; they are not the source of copied questions. Independent educator review is pending.
Explore analytical data processing - Training | Microsoft LearnWhat is Microsoft Fabric - Microsoft Fabric | Microsoft LearnExplore file storage - Training | Microsoft Learn