Expressions
Actions do not receive arguments from each other directly. Each one writes its result into the run's context under its own name, and later actions read from that context with an expression.
An expression is written between double braces:
{{['ActionName']}}
Reading an action's result
Every action produces a result, available under the action's name. Add a JSONPath expression to reach inside it.
{
"actionType": "Download",
"name": "DownloadFile",
"next": "UploadFile",
"url": "{{['GetProjectFile']$.Content.Uri}}"
}
$ is the root of the referenced result. .Content is where most actions put what they produced, so {{['ActionName']$.Content}} is the shape you will write most often.
Reading the triggering request
{{['#HttpRequest']}} holds the request that started the run: scheme, host, path, method, query, headers and body. It is the one context entry that is not an action.
{{['#HttpRequest'].Body.Name}}
{{['#HttpRequest'].Query.TenantName}}
Scheduled runs have no inbound request, so there is nothing useful in it. See How automations start.
Reading parameters
SetParameter stores a value under a name you choose, and it is read back exactly like an action result:
{{['SupplierCode']}}
Use it to compute something once and quote it in several later actions, rather than repeating the same long expression.
Preprocessors
#eval(<expression>) evaluates its argument first and substitutes the value, which lets you build an expression whose shape depends on data. It can appear anywhere in an expression, including inside a JSONPath filter.
{
"example1": "{{['#eval(['MapCustomDatabaseForm'].Name)'].Content}}",
"example2": "{{['RecordsList'].Content..#eval(['MapCustomDatabaseForm'].Id)}}",
"example3": "{{['RecordsList'].Content..#eval(['MapCustomDatabaseForm'].name).#eval(['MapCustomDatabaseForm'].Id)}}"
}
Reach for it only when a plain expression cannot express what you need. Nested preprocessors are hard to read and harder to debug.