The request

From Ethan Cole, Senior BI Developer
Your goal
Pick the change that removes the extra date hierarchies from this file and keeps them out of new files.
What Ethan gets, and why
The lesson
A fresh install of Power BI Desktop is set up for someone building their first chart in five minutes. Report builders change a handful of settings on day one, because the defaults make models bigger, guess things you’d rather decide, and read some files wrongly.
Where the options are
File → Options and settings → Options. The list on the left has two groups:
- Global: applies to Power BI Desktop on this computer and to every new file.
- Current File: saved inside the open
.pbix, so it travels with the file to colleagues.
Several options appear in both groups. Changing the Global one doesn’t touch files that already exist; open them and change Current File too.
Auto date/time: off
Global → Data Load → Auto date/time for new files, and Current File → Data Load → Auto date/time. When it’s on, Power BI creates a hidden date table for every date column in the model and adds a Year / Quarter / Month / Day hierarchy under it. With four date columns that’s four hidden tables, each covering every year its column touches. The hierarchies also behave differently from a real date table, which confuses time intelligence later.
The better habit, which this track uses from the start: one ‘Date’ table, marked as the date table (Table tools → Mark as date table), related to the date columns that need it.
Regional settings: read the data the way it was written
Current File → Regional Settings → Locale for import decides how text becomes dates and
numbers when data is loaded. A supplier file written in Romania says 03.04.2025 and 12,50; read
with an English (United States) locale, the date can come out as March 4 instead of April 3, and
the number may fail to convert. Set the locale to match the source (or set it per step in Power
Query with Change Type → Using Locale). Under Global → Regional Settings you also choose
the language of the Power BI interface.
Relationships: decide them yourself
Current File → Data Load → Relationships has Autodetect new relationships after data is
loaded. Power BI then guesses relationships from matching column names, and it sometimes
guesses wrong, for example joining two tables on a column called id that means different
things in each. Many modellers switch it off and draw every relationship themselves in Model
view, where they can check the cardinality and direction.
Default summarization: per column
Numeric columns are summed by default when you drop them into a visual. That’s right for
quantity, wrong for order_id, customer_id or a year number. There is no single switch for
this: select the column, then Column tools → Summarization → Don’t summarize (or set
Summarize by in the column’s properties in Model view). Do it for every ID and code column as
you build the model.
Two more on day one
- View → Themes: one theme for the whole report, so colours stay consistent.
- View → Gridlines and Snap to grid: tidy alignment.
About this level
Power BI Desktop runs only on Windows, so this quiz names the options instead of showing them. The practice levels in this track already follow these settings: one marked ‘Date’ table and no auto date/time.
Common mistakes
- Changing Global and expecting the open file to change.
- Trusting autodetected relationships without checking them in Model view.
- Leaving IDs summed: a “Sum of order_id” in a visual is a sign nobody set summarization.
Lantern lit!
Practised: optionsauto date timeregional settingslocale for importautodetect relationshipssummarization
- Solved
- No hints
- First run
