It would be nice if the PythonCaller had a parameter for a group-by so I don't have to code the logic each time I need it.
Released
Group-By parameter on the PythonCaller
This post is closed to further activity.
It may be an old question, an answered question, an implemented idea, or a notification-only post.
Please check post dates before relying on any information in a question or answer.
For follow-up or related questions, please post a new question or idea.
If there is a genuine update to be made, please contact us and request that the post is reopened.
It may be an old question, an answered question, an implemented idea, or a notification-only post.
Please check post dates before relying on any information in a question or answer.
For follow-up or related questions, please post a new question or idea.
If there is a genuine update to be made, please contact us and request that the post is reopened.
+40- Contributor
The code gets very interesting when I'm dealing with a custom transformer where the group-by is a user parameter, since the python api does not gracefully handle them. [see
http://knowledge.safe.com/idea/21410/python-caller-user-parameters-automatically-prefix.html]
+7I really like this idea if also the option 'Input is ordered by group' is present so that the PythonCaller doesn't become a blocker transformer when the input is ordered.
+22- Safer
As a workaround for this, if anyone is interested, you can stick you're PythonCaller inside a custom transformer. Custom Transformers have the option to enable groups and parallel processing. When inside a custom transformer, take a look at the Advanced Transformer Parameters in the Navigator Pane.
+40- Contributor
The issue with a custom transformer wrapper is that working with parameters becomes difficult, as the parameter name is prefixed by the transformer name, and that transformer name is variable.
The workaround to that is to use a parameterFetcher and then in python access the values as attributes rather than parameters. However that creates unnecessary attributes that then require cleanup, and risks overwriting actual attribute values if there is a name clash.
+49- Evangelist
How do you envisage this working in practice? Running the entire code for each set of features?
+40- Contributor
Yes, I would expect the entire code, including input and close methods to be run for each group of features.