Is there a way to get a random sample?

[argenisleon]

Looking at the docs, it seems there is no function to get a random sample of a dataframe.
Any way workaround to accomplish this?

[oleksiyskononenko]

Well, that's just a matter of generating a list of random indices from 0 to nrows - 1 and then using this list in the [i, j, ...] selector as DT[list_of_random_indices, :]. The list_of_random_indices can be easily generated by using the standard Python module random. For instance, random.sample() could be used for this purpose: https://docs.python.org/3/library/random.html#random.sample

If the random samples are needed for K-fold cross validation, then please take a look at the parallel kfold_random() function, that is a part of the datatable.models: https://datatable.readthedocs.io/en/latest/api/models/kfold_random.html