I'm sorry, but the IgniteDataset can not be used like a basic term, due to it's ML specific term and also we have a few kind of datasets (that are not equivalents of IgniteCache/Space and etc) like IgniteDataset, PartitionBasedDataset, LocalCachedDataset and so on.
Of course, we could find a dataset as a primitive in Spark (but the main reason to add this term there was a collaboration with Python - oriented data scientists) The real 'dataset' concept includes a lot of operations under the 'datasets': different math functions and etc My suggestion: IDataBucket -- Sent from: http://apache-ignite-developers.2346864.n4.nabble.com/