Free Databricks Certified Data Analyst Associate Actual Exam Questions

It’s B, refresh schedules are set outside the Query Editor.

B vs E? You definitely can’t set refresh schedules in Query Editor.

Heatmaps (A) mainly show intensity, like where users click most often, but they don’t depict the actual path users take across pages. Sankey (E) really fits when you want to see how users move from one page to another, showing the flow clearly. Choropleth and word clouds are off-topic here since they deal with location data and text frequency, respectively. Pivot tables are great for summaries, but not for visual flow. So if the goal is to track movement through the site steps, isn’t Sankey the only one that makes sense?

E imo. Sankey is all about flows between points, which fits tracking user paths through pages way better than heatmaps or word clouds, which are more static or focus on different data types.

A seems off since data quality testing usually needs data engineering tools, not just SQL. E looks like orchestration, so probably not Databricks SQL either. Does that narrow it down?

Makes sense to pick C since Databricks SQL is designed for querying and merging data sources directly. Tasks like data quality checks or ML tracking usually need different parts of the platform. C

This definitely seems like option C. External tables usually keep the data separate from the metadata, so dropping the table only removes the metadata, leaving the data files intact. If it were a managed table (E), both would disappear. The size or location of the table (A, B, D) shouldn't affect this behavior here.

C imo, external tables keep data separate from metadata, so dropping just removes the metadata but leaves data files intact. Managed tables would delete both.

Option E definitely makes the most sense here since extreme outliers skew the mean way more than the median, creating a noticeable difference. Also, A, B, C, and D don't really fit because missing values or variable types like boolean or categorical don’t impact the difference between mean and median in a meaningful way—mean and median aren't really applicable for non-numeric data. So it’s safe to say outliers are the key factor here.

Makes sense that outliers push the mean away from the median, so E stands out. Without extreme values, mean and median stay close, so A can be dismissed. E

B imo, since Data Explorer focuses on metadata and data access, not on dashboards or visualizations.

Maybe B is the best fit here since Data Explorer definitely lets you browse metadata and data. I’m not 100% sure about changing permissions either, but it seems more likely than the other options. A sounds off because UPDATE queries are usually done in SQL editors, not the explorer. C and D feel more about dashboards and visualizations, which are different tools. E is unrelated. So, B feels like the most reasonable choice even if "change permissions" might be a bit of a stretch.

C/B? This looks like an attempt to insert all rows from one table into another, but the use of TABLE after INSERT INTO isn’t standard SQL. Normally, you’d do INSERT INTO suppliers SELECT * FROM new_suppliers. So I’m thinking the command won’t run as is, which points to B. But if it did work, it would just add all rows including duplicates, so that matches C’s description. Still, syntax is the bigger problem here.

It’s B. The syntax isn’t right for any standard SQL I know—INSERT INTO usually needs VALUES or a SELECT statement, not a TABLE keyword like this. So it seems like this command would just throw an error rather than doing anything to the tables.

Option B, DROP TABLE is the standard way to remove just one table and its data files.

B/C? B seems legit for dropping the table and files. C looks off because DELETE TABLE isn’t a usual command, but maybe it’s a trick? Still, B is safer.

Maybe A makes the most sense here since it’s a small file, and the Create page is designed to handle straightforward uploads quickly without extra steps like cloud storage.

C imo makes sense too since uploading to cloud storage then importing is pretty standard and reliable. Even though it’s a bit more steps than A, it avoids any limitations the Create page might have with certain CSV formats or bigger files. Plus, once in cloud storage, you can reuse the data elsewhere easily.

Option D has typos but uses GROUP BY, which is necessary for aggregation.

It’s A for me too. Aside from the typo in D, you need the GROUP BY clause on customer_name to aggregate orders properly, which B lacks and C doesn’t even use count correctly. A’s syntax is clean and matches the expected output perfectly.

E imo, because you can't really ignore any of those points. Internal best practices are there for a reason, but legal requirements from both where the data’s collected and where it’s analyzed can come into play, especially with cross-border stuff. Saying just one misses the bigger picture.

Maybe A, since internal policies usually cover all legal bases anyway.

Thinking about it, A talks about inferring properties, which is more inferential stats, so that's not it. C narrows to quantitative only, but descriptive stats also summarize categorical data, so B feels right here. B

Option B looks good here because descriptive statistics isn’t just about numbers like means or SDs—it also covers categorical summaries like frequencies or modes. C narrows it down too much by focusing only on quantitative descriptions. Also, D talks about quantitative variables but that’s more about data types, not descriptive statistics as a whole. A is really about inferential stats, so that’s out. So B fits better since it includes all categorical summary stats, which are definitely part of descriptive statistics.

B vs C? I’m ruling out A since “never stops changing” is too vague and kinda weird wording. D is clearly off because it talks about categories, not continuous values. Between B and C, B talks about finite or countable values, which fits discrete variables better. Continuous variables should be able to take any value within a range — so uncountable sets, like in C. So C makes more sense here because continuous means you can have any decimal or fraction between numbers, not just countable points.

B tbh isn’t right because it says finite or countably infinite, which sounds more like discrete variables. D is about categorical variables, so that’s out. A is kinda vague and “never stops changing” doesn’t really capture the math behind continuous variables. C nails it with the idea of uncountable values in a range, which fits the definition of continuous variables best.

I’d say it’s C too, but for a slightly different reason: higher-order functions like map and reduce are designed to handle operations over collections or arrays efficiently, especially when dealing with big data frameworks. Options A and B feel off since they don’t really capture the “at scale” aspect or the array focus. D and E seem unrelated to when you’d specifically pick higher-order functions over other tools. So, applying custom logic at scale on array-like data fits best with C.

It’s C, because higher-order functions excel at handling operations on arrays and collections at scale.

Option D makes sense because business analysts focus on writing SQL queries to extract insights, unlike data scientists or engineers who use more code-heavy tools. SQL is their go-to.

D/C? Business analysts use SQL a lot, but platform architects might also work with Databricks SQL to set up query environments. Still, business analysts rely on it as their main tool more consistently.