Study guide

Study guide for chapter 6 of the Data 360 Architect course. Read the five modules first; this page helps you check what stuck.

About this study guide. It’s built only from this course’s own content and the published Salesforce exam guide’s list of objectives. It is never based on real exam questions, and it doesn’t claim to know what’s on the exam. If you’ve taken the exam, please don’t send me questions from it: the certification agreement forbids sharing them, and memorizing answers is the wrong way to become an architect.

This course covers the Data 360 Consultant exam, but the exam is just a step on the road. As architects, we go beyond it. Chapter 6 is what happens once data is through the door: the stream and its lake object, the category, the rows that fail, and the fields Data 360 adds that you never sent.

For the Consultant Exam

What this chapter covers of the Data 360 Consultant exam guide (opens in a new tab)’s objectives. Tick them off as you go. Each one links to where it’s taught.

ING-2: Bringing data in from many sources

• I know the three layers: data streams feed in, data lake objects hold the data as it came, data model objects are the shared model. (Three layers)
• I know the design work happens at the stream: source, fields, types, key, category, name, refresh. (Where the design work happens)
• I know renaming a stream doesn’t rename its lake object, fields are added at the stream, and lake object fields can’t be deleted. (What you can and can’t change)
• I know a lake object takes data from at most one stream, and can receive data from one or more transforms. (One lake object, one stream)
• I can choose the category: profile (about a person or account), engagement (a person’s activity at a moment), or other (reference data). (Three categories, in plain words)
• I know an engagement stream needs an event time field that never changes and can’t be edited after setup, and that a time series without one goes in as other. (Engagement needs an event time)
• I know a data model object inherits its category from the first lake object mapped to it, and engagement lake objects map only to engagement model objects. (What the category decides later)
• I know Data 360 adds KQ_ key qualifier fields beside keys (empty by default), lineage fields, and system fields to every lake object. (What Data 360 Adds)

SA-4: Diagnosing and troubleshooting

• I know failed rows land in a problem records lake object, named PR_ plus the lake object’s name, and are kept for 30 days. (They won’t just go away quietly)
• I start from a problem record’s error code: an empty key, a date that won’t parse, two rows with the same key. (What a problem record tells you)
• I can check a first load: Refresh History, the PR object, a sample in Data Explorer, and the key. (Check the first load)
• I know Refresh History keeps a stream’s last 50 refreshes, and a record-triggered flow on the Data Stream object can tell someone when a run fails. (Watch every load after)
• I know joins on qualified keys must use the key and its qualifier, null-safe, with COALESCE(). (Every join changes)

As an Architect

The consultant exam is the floor. This is what an architect takes from this chapter, beyond it.

• A lake object is a real table, found by name by everything built on it. Name it once. (Why it matters in this chapter)
• Watch the default category. File streams start as Profile; change it before you save, because it can’t be changed after. (Three categories)
• Flights are other data in this course: a flight isn’t a person’s activity, and its schedule keeps changing. Engagement is a person’s activity, like a traveler’s booking. (Flights: our rule)
• A formula field fixes what the row can fix; anything needing other rows is a transform (chapter 7). (Formula fields)
• Test with real volumes. Ten rows prove the stream works, not the design. (Test it with real volumes)
• A flow tells you without looking. In production, someone gets told when a run fails. (Watch every load after)
• Name qualifiers for the source, not the field, and stay within 20 active qualifiers. The KQ_ field holds the qualifier’s API name. (Setting one up)
• Don’t build on undocumented system fields. Ask the producer for the columns you’ll need to trace a record. (The system fields)

Words to Know

Term	In one line
Data stream	The feed in from one source, where the design decisions are made
Data lake object	The stored raw layer: a real table in the lake
Profile	A category for data about a person or account
Engagement	A category for a person’s activity at a moment, with a fixed event time
Other	A category for reference data, such as airports, runways and flights
Event time field	An engagement stream’s date of the event, which must never change
Problem records	Rows that failed to load, kept in a PR_ lake object for 30 days
Key qualifier (KQ_ field)	The field beside a key that says which source it came from
Lineage fields	Data Source, Data Source Object and Internal Organization: where a row came from

Check Yourself

Short scenarios in the style of the course. Decide your answer before you open it.

1. A team renames a data stream to match a new naming standard. Is the lake object renamed too?

No. Renaming a stream doesn’t rename its lake object, and everything built on the lake object finds it by its name. Name both right the first time. Reread What you can and can’t change.

2. An engagement stream uses a claim's last-modified date as its event time. What goes wrong?

The date changes as the claim moves through review, and that duplicates records with the same key. Use the date that’s fixed when the record is created. Reread Engagement needs an event time.

3. You can't find the model object you expect when you map a lake object. What do you check first?

The category. The list is filtered by compatible categories, and engagement lake objects map only to engagement model objects. Reread What the category decides later.

4. A stream loaded 980 of 1,000 rows. Where are the other 20, and how long do you have?

In the problem records lake object (PR_ plus the name), each with an error code. They’re kept for 30 days. Reread They won’t just go away quietly.

5. Two sources both use the key value PLT-000238 for different people. What keeps them apart, and what must every join then do?

A key qualifier per source. Every join then uses the key and its qualifier, null-safe with COALESCE(), because two empty values don’t match in SQL. Reread Every join changes.

6. Someone suggests making the flights an engagement stream, with the scheduled departure as the event time. What's this course's answer?

Flights are other data: a flight isn’t a person’s activity, and schedules change, sometimes weeks ahead. Engagement belongs to a person’s activity with a fixed time, like a booking. Reread Flights: our rule.

Explore

• Salesforce Certified Data 360 Consultant Exam Guide (opens in a new tab), Salesforce Help: the official list of objectives and their weights. Always check the current version before you book the exam.