Dataset -> Analyze column fails predictably on big datasets

Thomas_K Registered Posts: 15 ✭✭✭✭

I have a semi-big dataset (9GB, ~25 Mio entries). I tried to get a first look at it by doing Dataset -> Clicking on any column -> Analyze -> on "whole data".

This starts the following process:

... which fails everytime. I tried storing the dataset on Hive and as a single local file. It always fails, in Hadoop it fails considerably faster. (~4h).

Is this a known problem? Any suggestions what to do about it? If there is a connection error, shouldn't the results be calculated on the server still? Do I have to maintain a connection from my PC to the DataIku server the whole time?

Also, a side note:The progress bar doesn't really do anything? It stays white until it fails.

Setup Info
      Help me…