Using Dimensions (Hierarchies)

Dimensions in PlaidCloud are hierarchies — tree structures that organize things like cost centers, products, accounts, geography, or time periods. They’re the scaffolding that allocations, dashboards, and reports use to slice and aggregate data.

A dimension can carry more than just parent-child relationships. Each node can hold properties, aliases, and values — so a cost center hierarchy can also tell you which currency each center reports in, what business unit it rolls up to in an alternate view, and what its operating budget is.

Dimensions are managed in the Dimensions tab within each project.

Main Hierarchy

Every dimension has a main hierarchy. The main hierarchy defines the complete set of leaf members — every leaf node anywhere in the dimension must appear here.

Think of the main hierarchy as the canonical, single-truth tree. Anything in the dimension is a member of the main hierarchy; the question is just where in the tree it sits.

Alternate (attribute) Hierarchies

Alternate hierarchies are different views of the leaves in the main hierarchy. They can pick a subset of leaves, group them differently, or use entirely different roll-ups.

Two common patterns:

• Subset view — pull a specific set of leaves into a focused tree for a specific report or allocation. The alternate inherits any changes to its members from the main.
• Different roll-up — same leaves, different parents. For example: the main hierarchy organizes cost centers by department; an alternate organizes the same cost centers by geography.

Note

Items in the main hierarchy carry attribute labels showing which alternate hierarchies they also belong to.

An alternate hierarchy regroups the same leaf members under a different parent tree — here the region rollup and a size rollup share the identical leaves, so one dimension supports several ways to aggregate.

Managing Dimensions

Creating

From the New button in the toolbar, choose New Dimension. Enter a name, a directory (for folder-style organization), and a descriptive memo.

Click Create — the dimension is ready immediately. You can also create one from a workflow using the Dimension Create step.

Creating an Allocation Dimension

An allocation dimension (also called an assignment dimension) is the special dimension an allocation step reads to decide how each source record is spread across its targets. Rather than build one node by node, use the guided builder to create a starting dimension in a single pass.

Where your workspace has the feature enabled and you hold permission to create dimensions, open the New menu and choose New Allocation Dimension (marked with a split icon). The builder opens on the first of four steps:

1. Name and tables — Give the dimension a name, then pick the source (values) table whose rows your rules filter and the driver table the amount is spread against.
2. Rule filters — Check the source columns that act as rule filters. A column that already has a dimension relationship is honored as-is; for one that doesn’t, you can attach a dimension so its values resolve to members.
3. Driver and targets — Designate the driver-name and driver-value columns on the driver table, and check the target columns the amount is spread across.
4. Rules — Fill a flat grid of rules. Each rule is a set of source-filter values plus the named driver to use for that combination.

Click Finish and PlaidCloud creates the whole dimension in one step. It appears in the dimensions tree badged with a split icon, marking it as an allocation dimension rather than an ordinary hierarchy.

The dimension is authored independently of any allocation, so a single allocation dimension can serve several steps. To use it, point an allocation step at it — see Configure an Allocation.

Deleting

Select the dimension, open the Actions menu, and choose Delete Dimension. This removes the dimension and all underlying data.

You can also delete from a workflow using the Dimension Delete step.

If you want to keep the dimension but reset its contents (clear all structure, values, aliases, properties, and alternate hierarchies), use Dimension Clear instead of delete.

Copying

Select the dimension, open Actions, and choose Copy Dimension. Specify a name for the copy and click Create Copy. The copy includes values, aliases, properties, and all alternate hierarchies.

Sorting

The dimension management area lets you move members up and down and change parents directly. For large hierarchies, doing this by hand gets tedious — use the Dimension Sort workflow step to sort programmatically. It’s a big time saver after data loads or major restructures.

Loading Dimensions

Loading a dimension means converting tabular source data into hierarchical structure. PlaidCloud supports two data shapes for loads:

• Parent-Child — Two columns: one for parent, one for child. Each row defines one edge in the tree. This works for arbitrarily deep, irregular hierarchies.
• Levels — One column per level. Row by row, each column tells you the parent at that level. Best for regular hierarchies with predictable depth (e.g., country → region → city).

Loads can also carry values, aliases, and properties alongside the structure. See the Dimension Load workflow step for the full set of options.

Node Properties

Toggle Properties on the hierarchy toolbar to show a column for each property defined on the dimension — a cost center’s currency, an account’s business unit, a product’s category.

Adding or Editing a Property

1. Click New Property on the toolbar
2. In the Property Configuration window, enter a Property Name, choose the Property Display (Text or Tag), and pick an Editor Type — the control the property uses for data entry, including Select (Dimension), Select (Fixed), Select (Table), Checkbox, Spinner, Date Chooser, and Numeric
3. If the Editor Type needs further setup, a Configuration section appears with an Edit Configuration button
4. Click Create (or Update, when editing) to save

To change an existing property’s setup, select its column and choose Edit Property Configuration from the toolbar. Delete Property removes a property entirely; Clear Property clears the selected entry, and Clear All Entries For Property clears it across every node.

Dimension-Type Properties and the Member Picker

Setting a property’s Editor Type to Select (Dimension) turns it into a lookup against another dimension’s hierarchy (or the same dimension’s) — for example, a cost center’s “Region” property that points at a Geography dimension. Clicking Edit Configuration opens the Dimension Configuration window:

• Dimension and Hierarchy — the dimension and hierarchy the property picks members from. Select Not Specified on Hierarchy to let the picker show every hierarchy in the dimension.
• Start Node (optional) — restrict the picker to a subtree instead of the whole hierarchy.
• Allow Multiple Selections — lets a single property value hold more than one member. It’s on by default; untick it to restrict the property to a single member.
• Special Cases — offers Source and Current, extra selectable entries alongside the hierarchy’s real members.

Save requires both a Dimension and a Hierarchy to be chosen.

Once a property is configured this way, double-click one of its cells in the grid to open the member picker — the same tree-style picker also appears for a dimension-mapped column in a Data Editor grid. It opens on the first double-click, whether or not the cell already has a value.

• Expand All / Collapse All — open or close every branch of the tree.
• Click a member (or, with Allow Multiple Selections on, several members) to select it.
• Apply sets the property to your current selection.
• Clear removes the property’s current value.
• Cancel closes the picker without changing anything.

Note

If several rows are selected in the grid when you open the picker, Apply sets the value on all of the selected rows, not just the one you double-clicked — this is intended for editing many rows at once.

Dimension Property Inheritance

A dimension can be configured so that children inherit property values from their ancestors. To turn this on, click the dropdown next to Properties and select Inherited Properties.

Notes on how inheritance behaves:

• Inheritance applies to all properties in the dimension — you can’t enable it for one property and not others.
• If you set a property on a child, then later delete that value, the child reverts to its parent’s value. Children cannot have a null property when the parent has a value.
• Setting a property on a node propagates down to its descendants, overriding their inherited value until they’re explicitly set.
• Inheritance walks the tree all the way down to leaf nodes.

Finding Nodes in a Hierarchy

A large dimension can run to thousands of members, so the hierarchy window has two ways to narrow what the grid shows. Both hide non-matching nodes rather than merely highlighting them, and both keep a matching node’s ancestors on screen so you can still see where it sits in the tree.

The Filter Box

The Filter: box at the bottom of the window filters as you type. It matches on any part of the text, ignoring case, and searches:

• the node’s name in the hierarchy, and
• every Alternates, Aliases, Properties and Values column you currently have showing.

Because it only searches the columns on screen, the toolbar toggles double as the search scope — turn Properties on to find a node by an account code held in one of its properties, and turn it off again to search names alone. A property displayed as tags matches on any single tag.

The buttons beside the box choose which pane the filter applies to: Both, Main Hierarchy, or Alternate Hierarchy. The funnel button clears the filter and restores the full tree.

Note

The Filter box searches the value stored on each node. With Inherited Properties turned on, a node showing a value it inherited from an ancestor is not matched by that value — only the ancestor the value is actually set on is.

Advanced Filters

Where the Filter box searches every showing column at once, Advanced on the hierarchy toolbar builds precise, per-column criteria. It opens the Configure Advanced Dimension Filters window, where each row is one criterion:

1. Click Add Row
2. Pick a Column — any property or aggregation column on the dimension
3. Pick a Condition. A property holds text, so it offers =, != and contains; an aggregation holds numbers, so it offers =, !=, <, >, <= and >=
4. Either type a Value to compare against, or choose another column under Compare With to compare the two columns against each other — one or the other, not both. Compare With lists the other columns holding the same kind of value, since comparing a number against text answers nothing either way
5. Click Update to apply

Delete Row removes the selected rows and Clear empties the whole list. Add several rows to narrow further: a node has to satisfy every row, and text in the Filter box as well. Update stays disabled until every row is complete, so a half-built criterion never silently changes what you see.

Advanced criteria apply to the main hierarchy. The Filter box still applies to whichever pane the buttons beside it select.

Text conditions ignore case, the same as the Filter box. = matches the whole value, so use contains to find a code sitting inside a longer property value or in one tag of a multi-tag property.