Custom Transformer Parameters
Name of the custom transformer
Determines how the transformer is referenced in workspaces. May depend on whether the transformer is exported or not.
If Yes, the transformer is no longer available for use in newly-created workspaces. It remains usable only in existing workspaces.
Enable Blocked Looping
If Yes, the transformer can bypass its blocking behavior when used in a loop, and the following additional parameters apply:
- Maximum number of iterations
- Atttribute to hold iteration count
For more information, see Looping with a Blocking Transformer.
Incoming Attribute Preservation
Use this parameter to optimize performance by controlling how workbench determines the output of attributes that enter the custom transformer. If you already know all incoming attributes either pass through or are dropped by the custom transformer, the All Attributes or Drop Attributes settings, respectively, allow workbench to avoid the calculations that are otherwise necessary to achieve the same results.
- Automatic: Workbench determines which incoming attributes are output by each port.
- All Attributes: Instructs workbench to pass through all incoming attributes to all output ports. Use this setting if you already know the custom transformer performs in this manner.
- Drop Attributes: Instructs workbench to eliminate all incoming attributes (that is, no incoming attributes are output). Use this setting if you already know the custom transformer performs in this manner, or you do not need workbench to provide a final display of output attributes.
Specifies the number of threads the custom transformer can process in parallel.
|Level||Number of Processes|
|Minimal||Cores / 2|
|Moderate||Exact number of cores|
|Aggressive||Cores * 1.5|
|Extreme||Cores * 2|
For more information, see "About Parallel Processing" in the FME Transformers Help.
Parallel Process By
Specifies an attribute or parameter by which to group parallel processing threads. For example, if the attribute 'STATE' is specified, the transformer processes each state in parallel.
Parallel Process Groups are Ordered
If Yes, the FME engine can assume that data is inputted to the custom transformer in order for each 'Parallel Process By' parameter value. This setting provides a performance and memory boost, because the engine does not need to wait for all features to arrive before beginning a process. Set this value to Yes only if the structure of your dataset allows for data throughput in order by the 'Parallel Process By' setting.