I have indexed a bunch of documents containing an optional field called value
. How do I use Eland to take a random sample of n non-null instances of value
with replacement?
If I had a pandas data frame I would do the following.
f.dropna().sample(10, replace=True)["value"]
However neither dropna
or the replace
argument of sample
are supported in Eland data frames.
Is there a way to do this with an Eland data frame? If not, how do I do it in Elastic? I'm looking at the random sample aggregator but it looks like that doesn't sample with replacement.