+14Hi everyone,
I'd like to create an arbitrary subset of features and can't find a suiteable way to do this. My features have a unique ID and I'd like to select a sample of n-features of it. How can I get n-individual random numbers to test these numbers against my ID and filter the features this way? Or how can I fetch n indices from a list that might be created from the features?
Here is some sample data:
Best answer by nielsgerrits
I use a Sampler for this, but there are probably better ways to randomize the sample.
+60I use a Sampler for this, but there are probably better ways to randomize the sample.
I use a Sampler for this, but there are probably better ways to randomize the sample.
Agree that the Sampler is the way to go. The modern Sampler has the Randomize Sampling parameter.
+14I use a Sampler for this, but there are probably better ways to randomize the sample.
Yes, the sampler does the trick, although it is not possible to hand over the sampling rate to an attribute created during runtime. In case the rate exceeds the feature count, the sampler can handle this situation without any problems and routes all input features to the sampled port.
