Create a dimension in Business Context

A dimension is what you break a number down by — product, region, month, price band. Measures give you the figures; dimensions give you the ways of slicing them.

There are three kinds, and they're built differently, so this article covers each in turn.

Before you start: open a semantic layer and go to Business Context → Dimensions.

Every kind starts the same way — click Create and pick one.

Tip: before building from scratch, try Suggestions — Astrato reads your model and proposes dimensions from the columns already there. See Get suggestions from your data.

The straightforward one: a column from your model, exposed as something people can group by.

Step 1: Choose Create → Field dimension.

Step 2: Give it a Name — this is what people will see, so use the business word rather than the column name.

Step 3: Add a Description if it needs one.

Step 4: Under Field binding, choose the column it comes from. The list is grouped by table, with a search box for larger models, and each entry shows its data type.

Step 5: Choose a Format if the values need one — dates and numbers usually do. It appears as a chip you can remove again.

Step 6: Click Create.

A drill-through is a path — a set of dimensions in order, so someone can start at the top level and step down into detail. Region to country to city, say, or category to subcategory to product.

Step 1: Choose Create → Drill-through.

Step 2: Name it for the journey rather than any one level — Product hierarchy rather than Category.

Step 3: Click Add and pick the first level.

Step 4: Keep adding levels. The order matters — it's the order people will travel through, from broadest to narrowest.

Step 5: Click Create.

A bucket groups a continuous number into bands: age groups, price brackets, order-size tiers. It turns a column of numbers into a handful of readable categories.

Step 1: Choose Create → Bucket.

Step 2: Name it and describe it.

Step 3: Pick the Source field — the number you want banded.

Step 4: Choose a Format for how the band labels appear. Each option shows you a live example, so you can see the result before choosing. No format uses the raw value, and Custom format is there if none of the presets fit.

Step 5: Set the Distribution — the Min value and Max value the bands should span.

Check your bands before you commit. Use Show distribution to see how many rows land in each band. If nearly everything piles into the first bar, your ranges don't match your data and the dimension won't tell anyone much.

Step 6: Click Create.

In Settings → Grouped dimensions, choose what the Base field picker offers:

Both options are off by default and are saved for the semantic layer. Turning the first option off also clears the second.

Only dimensions backed by a single existing, non-calculated field without a date transformation are offered. Buckets, drill-throughs and other custom groups cannot be used as the base field.

Choosing a dimension stores its underlying table field. The closed picker shows that table and field, so two dimensions backed by the same field produce the same selection. Renaming or deleting the dimension you chose does not change the group's field binding. These settings control the picker options; they do not restrict document editing or agent authoring.

A sets dimension groups an answer by stored set filters: each value is one set, such as "High spend" or "Repeat buyers", and a member of several sets appears under each. In Dimensions, select Add → Sets, then choose the set filters under Add and, optionally, a label for each. Set filters must be created first, in Business Context → Set Filters.

Use a sets dimension like any dimension in the Data Explorer and in Business Questions. It is available where Business Questions are, with these limits:

Deleting a set filter removes it from the sets dimensions that use it; a sets dimension left with no sets is deleted.

Your dimension appears in the list with its source underneath — here 1 Sales › Orderdatekey — along with when it was last changed and by whom.

The toolbar above works like the one on Measures: search by name, filter by type, and Group by type when the list grows.

If a dimension has a display label, the list shows the label in place of its name. Search matches both, and hovering the row shows the label with the name beside it.

The ⋮ menu on each row offers the same four actions as a measure:

Show in lineage is the useful one. It opens the graph with your dimension highlighted and everything unrelated faded out, so you can see at a glance which table and field it rests on. Clear returns the graph to normal.

Once published, a dimension is also a filter subject in Data Explorer. Open Filter, click Add filter, and the field picker lists your dimensions first, under Dimensions, ahead of the table columns.

Pick one and the value list fills from the dimension itself: a field dimension lists the values of the column it reads, and a bucket or grouped dimension lists the bands or groups it computes, so what you pick is what the filter matches. A drill-through spans several columns and has no single value list, so its value is typed instead.

Nothing is created until you click Create, and Cancel backs out cleanly.

A new dimension goes into your working copy. It doesn't reach anyone else's workbooks until you publish, and it's undoable through version history — see How versioning works in the semantic layer.

Tip: name a version before adding several dimensions at once. It gives you one clean point to return to if the set isn't what you wanted.