Overview
Importing Data is the smallest section, tied at 5% with data modeling, and has two objectives. The first is conceptual: explain the approaches for getting data into the platform. These are reading files from Amazon S3 or other cloud object storage through Unity Catalog external locations, receiving data from other organisations through Delta Sharing, pulling data in through APIs, incremental file ingestion with Auto Loader, managed connectors in Lakeflow Connect for SaaS applications and databases, and installing Marketplace listings. Lakeflow Connect is new to the October 2026 wording.
The second objective is practical: upload a local file, such as a CSV or JSON file, through the workspace UI to create a table or place it in a volume. Analysts are not expected to build engineering pipelines, but they are expected to know which approach suits a scenario and what each one does with schema, freshness and copying.
How it's tested
Stems describe where the data is and how often it changes, then list several ingestion methods that could plausibly work. Choose the one that fits with the least custom code and the right freshness. Watch for words like incremental, one-time, partner, SaaS and no copy.
-
Pick the ingestion path
A source such as an S3 folder, a Salesforce org or a partner's lakehouse is described, and you choose the matching ingestion method.
-
Incremental files
New files land continuously in cloud storage and must be processed once each, pointing to Auto Loader rather than repeated full loads.
-
Quick upload
A business user has a small local spreadsheet to analyse today, and the answer is the workspace UI upload rather than a pipeline.
A supplier already stores inventory data in Delta tables on its own Databricks account and wants to give your team live read access without sending files. Building an S3 export plus Auto Loader would work but adds copies and delay. Delta Sharing lets the supplier share the tables directly, and your team mounts the share as a catalog and queries current data in place, so Delta Sharing is the best answer.
Objectives 2
Official objectives from the Databricks outline, in outline order.
3.1 Explain the approaches for bringing data into Databricks, covering ingestion from S3, data sharing with external systems via Delta Sharing, API-driven data intake, Auto Loader, Lakeflow Connect managed connectors, and Marketplace.The main ingestion options and when each fits.
- What it means
- Files in Amazon S3, ADLS or GCS are read through Unity Catalog external locations and volumes with commands such as COPY INTO, read_files or CREATE TABLE AS SELECT. Delta Sharing is an open protocol for sharing live tables across organisations and platforms. API-driven intake pulls data from REST endpoints or pushes events in. Auto Loader incrementally processes new files as they arrive. Lakeflow Connect supplies managed connectors for SaaS apps and databases. Marketplace offers ready-made data products delivered by sharing.
- What to understand
- Auto Loader tracks which files it has processed, supports schema inference and evolution, and suits continuous or scheduled incremental loads. COPY INTO is idempotent, skipping files it has already loaded, and suits simpler batch loads. Lakeflow Connect connectors handle authentication, incremental change capture and scheduling for sources such as Salesforce, Workday, ServiceNow and SQL Server. Delta Sharing and Marketplace deliver data as read-only catalogs without copying it. API intake needs custom code or partner tools and is used when no connector exists.
- How it shows up
- A source and freshness requirement are described; choose the method with the least custom work.
- Common confusion
- Choosing Auto Loader for a SaaS application, when it only reads files from cloud storage.
3.2 Use the Databricks Workspace UI to upload a data file to the platform.Creating a table or volume file from a local upload.
- What it means
- The workspace lets you upload files such as CSV, TSV, JSON, Avro, Parquet, ORC or text through the add data or create table page. You pick a target catalog and schema, preview the parsed data, adjust column names, types and header settings, then create a managed Delta table. Files can also be uploaded into a Unity Catalog volume.
- What to understand
- Uploading is meant for small, one-off files, not recurring feeds. The result is a managed Delta table governed by Unity Catalog, so you need CREATE TABLE on the schema plus USE CATALOG and USE SCHEMA. Check inferred types during preview, because a numeric column with stray text may be read as a string. Uploading to a volume keeps the raw file for later reading with SQL functions such as read_files.
- How it shows up
- A user with a local spreadsheet needs it in Databricks quickly; the UI upload is the expected answer.
- Common confusion
- Building an Auto Loader stream for a single small spreadsheet.
High-yield concepts
The terms, rules and formulas this domain turns into questions most often.
-
External location
Unity Catalog object pairing a cloud path with a storage credential.
- Remember
- Governs access to S3 or other object storage.
- Tested as
- S3 ingestion questions.
- Trap
- Using raw keys in queries.
-
Auto Loader
Incremental file ingestion from cloud storage.
- Remember
- Processes each new file once, infers and evolves schema.
- Tested as
- Continuous file arrival.
- Trap
- Applying it to SaaS or databases.
-
COPY INTO
SQL command to load files into a Delta table.
- Remember
- Idempotent: already-loaded files are skipped.
- Tested as
- Simple batch loads.
- Trap
- Expecting duplicates on rerun.
-
read_files
SQL table-valued function for reading files.
- Remember
- Works with paths in volumes or locations.
- Tested as
- Ad hoc file queries.
- Trap
- Thinking it creates a table on its own.
-
Lakeflow Connect
Managed connectors for SaaS apps and databases.
- Remember
- Handles auth and incremental capture.
- Tested as
- New to the 2026 guide.
- Trap
- Writing custom API code instead.
-
Delta Sharing
Open protocol for live data sharing.
- Remember
- No copy; recipient queries in place.
- Tested as
- Partner data scenarios.
- Trap
- Assuming files are exported.
-
Marketplace listing
Published data product installed as a catalog.
- Remember
- Delivered by Delta Sharing.
- Tested as
- Third-party data scenarios.
- Trap
- Expecting a write-capable copy.
-
API-driven intake
Ingesting from REST endpoints with code or tools.
- Remember
- Used when no managed connector exists.
- Tested as
- Custom source scenarios.
- Trap
- Choosing it when a connector exists.
-
UI file upload
Create a managed table from a local file.
- Remember
- Best for small, one-time data.
- Tested as
- Quick analysis scenarios.
- Trap
- Using it for recurring feeds.
-
Volume
Unity Catalog object for non-tabular files.
- Remember
- Holds uploaded raw files.
- Tested as
- Upload destination questions.
- Trap
- Treating it as a table.
Side by side
Concepts candidates mix up, compared on the point that decides the answer.
Choosing an ingestion approach
| Concept | Meaning | Key difference | Common trap |
|---|---|---|---|
| Auto Loader | Incremental processing of new files in cloud storage. | File-based, continuous or scheduled. | Used for database or SaaS sources. |
| Lakeflow Connect | Managed connectors to SaaS and databases. | No custom extraction code. | Overlooked in favour of APIs. |
| Delta Sharing | Live, read-only access to another party's tables. | Data stays with the provider. | Thought to copy data. |
| UI upload | Local file to managed table. | Manual and one-time. | Proposed for daily feeds. |
Batch file loading options
| Concept | Meaning | Key difference | Common trap |
|---|---|---|---|
| COPY INTO | Loads new files into an existing Delta table. | Skips files already loaded. | Assumed to reload everything. |
| CREATE TABLE AS SELECT with read_files | Creates a table from file contents in one step. | Full read each time it runs. | Rerun for incremental needs. |
| Auto Loader | Streaming-style incremental ingestion. | Scales to millions of files with checkpointing. | Thought to be SQL only. |
Exam traps
Each trap is an answer that looks right. Learn the clue that should change your mind.
-
Auto Loader for SaaS
- What candidates pick
- Auto Loader to ingest Salesforce opportunities.
- Why it looks right
- It is the best-known ingestion feature.
- Why it's wrong
- Auto Loader only watches cloud storage paths for files; a SaaS source needs a Lakeflow Connect connector or API intake.
Clue to watch for A named SaaS application or database source.
-
Copying shared data
- What candidates pick
- Export the partner data to S3, then load it.
- Why it looks right
- Files are familiar.
- Why it's wrong
- It adds latency, storage and pipelines when the partner can share live tables via Delta Sharing.
Clue to watch for Partner already uses a lakehouse or Delta tables.
-
Pipeline for a spreadsheet
- What candidates pick
- Set up Auto Loader for one CSV.
- Why it looks right
- It sounds robust.
- Why it's wrong
- For a single small file, the workspace upload creates a managed table in minutes with no code.
Clue to watch for One-time, small, local file.
-
Reloading with COPY INTO
- What candidates pick
- Rerunning COPY INTO duplicates rows.
- Why it looks right
- Many load tools duplicate on rerun.
- Why it's wrong
- COPY INTO tracks loaded files and skips them, so reruns are safe for unchanged files.
Clue to watch for The same command runs on a schedule.
-
Custom API code when a connector exists
- What candidates pick
- Write Python to call the source's REST API.
- Why it looks right
- APIs work for anything.
- Why it's wrong
- Managed connectors already handle authentication, incremental changes and schedules, so custom code adds maintenance.
Clue to watch for A supported source is named.
Answer framework
Use five questions to choose an ingestion method from the scenario details.
- Identify the source type: local file, files in cloud storage, SaaS application, database, partner lakehouse or public data product.
- Ask whether the load is one-time or recurring, because one-time small files point to the workspace UI upload. Uploaded CSV or JSON files become a Delta table in a schema you choose.
- For recurring files in S3 or other object storage, choose Auto Loader for incremental arrival or COPY INTO for simple batches.
- For SaaS applications and databases, check for a Lakeflow Connect connector before considering custom API intake. Managed connectors handle schema changes and incremental pulls for you.
- If the data already lives with another organisation or in Marketplace, prefer Delta Sharing so it is queried in place.
Practice question
Want more like this? The Databricks DA Assoc practice test has questions for every domain.
Checklist
0 of 10
Tick each one you can do without notes. Your ticks stay in this browser.
Study strategy
This section is small, so aim for clear decision rules rather than depth. Build a table with one row per approach and columns for source type, freshness, whether data is copied, and the typical exam keyword. Six rows are enough: S3 files with COPY INTO or read_files, Auto Loader, Lakeflow Connect, API intake, Delta Sharing and Marketplace, plus the UI upload.
Then do two short labs. First, upload a small CSV through the workspace UI into a schema you own, change one column type in the preview, and look at the resulting managed table in Catalog Explorer. Upload the same file into a volume and query it with read_files to compare the two paths. Second, read the Auto Loader overview in the documentation and run its quickstart notebook if your workspace allows, paying attention to the checkpoint and schema location.
Read the Lakeflow Connect overview to learn which kinds of sources it supports, since this wording is new to the October 2026 guide. Finish by writing three scenario questions of your own and answering them with your table.
FAQ
How many questions come from Importing Data on the Databricks Data Analyst exam?
Importing Data is 5% of the exam, about two of the 45 scored questions. It is one of the two smallest sections.
Do analysts need to write Auto Loader code?
The objective asks you to explain approaches, so focus on what Auto Loader does and when to use it rather than on syntax details.
What is Lakeflow Connect?
Lakeflow Connect provides managed connectors that ingest from SaaS applications and databases with built-in incremental capture. It was added to this objective in the October 2026 guide.
Does Delta Sharing copy data into my account?
No. Recipients query the provider's data in place through a read-only catalog, so the data stays current.
Which file types can I upload through the UI?
Common formats include CSV, TSV, JSON, Avro, Parquet, ORC and text. Uploads are meant for small, one-off files.
Is S3 the only cloud storage tested?
The objective names S3, but the same pattern of external locations and volumes applies to ADLS and Google Cloud Storage.
Other Data & Analytics certifications
Where the same subject shows up in other exams. Cards marked "Same subject" open the matching domain.
Test your Domain 3 knowledge
Practise choosing between Auto Loader, COPY INTO, Lakeflow Connect, Delta Sharing and the workspace UI upload under exam time pressure.
Keep going
- Training Training Built on your Weak areas
- Practice test Questions by domain, with explanations
- Mock exam Full length, timed like the real exam
- Readiness test Find your weakest domain in minutes
- Cheat sheet Key terms and formulas on one page
- Ask the Databricks DA Assoc AI tutor Get any concept explained your way