Transform nodes like Site Fill bundle several settings into one object: a building width, a minimum space between buildings, and so on. Sometimes you only want to drive one of those settings from elsewhere in your flow, like a project-wide slider, without rebuilding the whole settings object by hand or exposing every input separately. Write Property lets you reach into that one setting and replace it, leaving the rest as they are.
This works the same way on any dendrogram-category transform node, including:
and others. Not every dendrogram node has this Transform input: Centroid, Enclosed Polygon, and Explode do not, so this technique does not apply to them.
What a transform node outputs
Nodes in the dendrogram category, like Site Fill, Grid Fill, and Round Corners, have a T connector that accepts a settings object. A paired transform node, like the Site Fill transform node, has no inputs of its own, only an output: it generates that settings object from whatever you set with Edit Defaults or in the Properties Palette.
Each transform node is built to pair with one specific core node. The Site Fill transform node outputs exactly the object the Site Fill node's T connector expects, and nothing else. It will not work as the transform for Grid Fill or any other node, since each core node expects its own type and its own set of keys inside parameters.
Here is what the Site Fill transform node outputs, and what the Site Fill node's T connector expects to receive:
{
"type": "siteFill",
"parameters": {
"width": 24,
"space": 3,
"hasInnerGrid": true
}
}This is a plain object, not a GeoJSON feature. It has no properties { } wrapper of its own; the settings sit directly inside parameters. Every dendrogram-category node follows this same shape, with its own type and its own set of keys inside parameters.
For the difference between Transform mode and Exposed Inputs mode, see Transform Nodes & Exposing Inputs. For Site Fill's own settings, see Site Fill.
Override one setting with Write Property
Add the transform node that pairs with your core node, like the Site Fill transform node, to the canvas. Set its starting values with Edit Defaults if you want a specific baseline. Then connect the transform node's own output into a Write Property node, instead of connecting it straight to the core node's T connector:
Object: the transform node's output.
Property: the path to the one setting you want to change, like
parameters.width.Value: a Number node, or any computed value from elsewhere in your flow, like an Input Parameter tied to a project-wide setting.
Connect Write Property's Object output to the core node's T connector, in place of the direct connection from the transform node. Write Property is the right node here, not Write Feature Property, because this settings object is not a feature. See Find a Property Path with the Reader Node for why those two nodes exist separately.
📷 Screenshot placeholder: the Site Fill transform node's output connected into a Write Property node's Object input, a Number node feeding the Value input, and Write Property's Object output connected on to the Site Fill node's T connector. Replace this block with the image.
Before Write Property | After Write Property |
{ | { |
Only width changed. space and hasInnerGrid stay exactly as they were in the starting object.
Confirm the exact key with the Reader before typing it into Property. A setting's label in the properties palette, like "Width," is not always spelled the same way as its key, like width.
Unpublish the setting you are overriding
Overriding a value through a flow always wins over whatever is published in the Properties Palette. Once Write Property's result feeds into the T connector, that computed value overrides anything a person types into the published field for that same setting.
🚨 Warning: If you leave the setting published in the Properties Palette, a person's edit there is silently overridden by the flow the next time it runs. The field looks editable, but the flow controls the real value.
The best use case for this technique is unpublishing the setting entirely: turn its toggle off in Edit Defaults so it never appears in the Properties Palette. That avoids showing a control that looks editable but does not actually work, since the flow already controls the value.
When Exposed Inputs is a better fit
Some transform-core pairs can switch to Exposed Inputs mode instead, which gives each setting its own connector rather than chaining several Write Property nodes together. Very complex pairs do not have an Exposed Inputs option at all, so Write Property may be the only way to drive one of their settings dynamically. Reach for Write Property whenever you only need to override one or two values while keeping the rest fixed, or when Exposed Inputs isn't available.
See Transform Nodes & Exposing Inputs for how to switch a node to Exposed Inputs mode. See Write Property for its full input and output reference.