[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: re: statsby slowness

From   "Nick Cox" <>
To   <>
Subject   st: RE: re: statsby slowness
Date   Sun, 19 Aug 2007 15:48:49 +0100

It would be interesting to learn how another 
approach compares: computing the ranks using 
-egen, rank()- and then computing the Spearman 
via -egen, corr()- from -egenmore- from SSC. 
You should only need to -sort- once. 


David Airey
> Just to follow up, I had complained about the slowness of statsby  
> when the number of "by groups" is large, as in gene microarray data.  
> In my case the number of by groups was about 22,000 and the rows  
> about 300,000. I wanted to run a spearman correlation, and 
> return the  
> results to a file using statsby. I chose to break the file into  
> smaller files of 500 "by groups" each. I then found statsby was  
> performing at a speed that I could tolerate. That speed was 11  
> minutes for the main loop of about 22,000 tests, (2000 test every  
> minute or .03 seconds per test). Since statsby is a very convenient  
> command, this seems a reasonable approach to recommend to others.

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index