These are chat archives for thunder-project/thunder

14th
Feb 2015
Jeremy Freeman
@freeman-lab
Feb 14 2015 02:08 UTC
@d-v-b if you were only going to do this once (and immediately call some output action, like mean()), then it doesn't really matter, though the first one would probably be faster because it avoids writing data
but if you are going to load the Series data again even once, the second approach is probably better, because the next time you load you avoid the shuffling that happens during toSeries()
you could cache the toSeries output in the first version, but you might not have enough RAM for that
if it's the raw data