A node specifying the execution of a user-defined function.
| Name | Type | Description | Notes |
|---|---|---|---|
| registeredUdfName | String | If set, the name of the registered UDF to execute, in the format `namespace/name`. Either this or `executable_code` should be set, but not both. | [optional] |
| executableCode | String | If set, the base64 serialization of the code for this step, encoded in a language-specific format (e.g. Pickle for Python, serialization for R). | [optional] |
| sourceText | String | Optionally, the source text for the code passed in `executable_code`. For reference only; only the code in `executable_code` is actually executed. This will be included in activity logs and may be useful for debugging. | [optional] |
| environment | TGUDFEnvironment | [optional] | |
| arguments | List<TGUDFArgument> | The arguments to a UDF function. This encompasses both named and positional arguments. The format is designed to provide compatibility across languages like Python which have a fairly traditional split between positional arguments and named arguments, and languages like R which has a rather unique way of specifying arguments. For Python (and most other languages), all positional arguments will come before all named arguments (if any are present): // fn(arg1, arg2, arg3) [ {value: arg1}, {value: arg2}, {value: arg3}, ] // fn(arg1, arg2, n=kw1, a=kw2) [ {value: arg1}, {value: arg2}, {name: "n", value: kw1}, {name: "a", value: kw2}, ] // fn(kw=k1, only=k2) [ {name: "kw", value: k1}, {name: "only", value: k2}, ] However, in R, named and positional arguments may be intermixed freely: // fn(arg, n=kw1, arg2) [ {value: arg}, {name: "n", value: kw1}, {value: arg2}, ] | [optional] |
| resultFormat | ResultFormat | [optional] |