Hey! All data types are always kept serialized for caching/hashing/sorting. Deserialization is sometimes needed in the internal algorithms (on hash collisions and sort-prefix collisions). The most efficient data types for that are actually Tuples. POJOs and other data types are a little less efficient for that, but we are trying to improve that.
Stephan On Thu, Feb 19, 2015 at 10:23 AM, Kruse, Sebastian <sebastian.kr...@hpi.de> wrote: > Hi everyone, > > > > I think that during one of the meetups, it was mentioned that Flink can in > some cases operate on serialized data. Given I understood that correctly, > which cases that would be, i.e, which data types and operators support such > a feature? > > > > Cheers, > > Sebastian > > > > > > --- > > Sebastian Kruse > > Doktorand am Fachbereich Information Systems Group > > Hasso-Plattner-Institut an der Universität Potsdam > Prof.-Dr.-Helmert-Str. 2-3, D-14482 Potsdam > Tel +49 331 5509 240 > > Amtsgericht Potsdam, HRB 12184 > Geschäftsführung: Prof. Dr. Christoph Meinel > > > > >
