h2o group_by simple speed test

h2o-group_by.txt

########### R

library(h2o)

h2oServer <- h2o.init(max_mem_size = "50g", nthreads = -1)

d <- h2o.importFile(h2oServer, path = "d.csv")

system.time({
  print( h2o.group_by(d, "x", sum("y")) )
})

##   user  system elapsed 
##  0.232   0.003   5.988 

TODO: top 5 (order by, limit 5)



########### python

import h2o

h2o_server = h2o.init(max_mem_size_GB = 50)

d = h2o.import_file("d.csv")

%time d.group_by("x").sum("y").get_frame()

##CPU times: user 11 ms, sys: 4.21 ms, total: 15.2 ms
##Wall time: 5.27 s

TODO: top 5 (order by, limit 5)



########### setup

3.2.0.9 
16 cores