[R] help on qcc

David Winsemius dwinsemius at comcast.net
Mon Dec 3 15:24:12 CET 2007

Thorsten Muehge <MUEHGE at de.ibm.com> wrote in
news:OF7566C5AF.C84AC5C2-ONC12573A6.0036ACCB-C12573A6.003D04FA at de.ibm.com

> Hello R Experts,
> I started to work with the qcc package and it wprks quite nicely.
> Heres My Code:
> n <- 
> c(55,5,94,25,10,15,15,40,44,34,90,114,204,37,30,28,12,68,64,29,24,14,31
> ,16,62,45,55,20,24,14,9,19,76,57,55,42,6, 
> 54,32,117,19,32,9,11,13,31,27,33,44,28) x <-
> c(6,0,30,5,2,1,4,5,2,9,12,24,64,3,9,21,9,48,15,4,3,0,1,0,6, 
> 7,2,0,3,2,0,5,1,2,6,9,1,6,2,0,1,4,8,0,1,1,3,0,0,0) 
> qcc(x,type="p",size=n,title = "p-Charts Failed data Transmission");
> I see the numbers beyond limits and the number of violating runs.
> How can I switch of the number of violating runs?
> How can I specify the number of violating runs?

I am not an experienced user of qcc, (nor am I an expert), but I doubt 
that very many of the other readers of r-help are users of qcc either. 
Looking at the code of qcc and then at help for the package, the number 
of violating runs is determined through the shewhart.rules() function and 
the gcc.options$("run.length") argument. It looks as though you could 
send arguments to the rules= parameter when calling qcc() that would 
override the default. Looking further at the package documantation and 
the qcc.options() function my suggestion (which I tested and seems to 
have the effect I believe you requested) would be to try something like:

qcc.options("run.length"=5) #before the qcc(...) invocation

....and if you wanted to "turn it off", then set "run.length" to 
something impossibly high.

qcc.options("run.length"=1000) # number of violating runs becomes zero

If that does not work, then another method for getting help with a 
package is to send an email to the package maintainer.

David Winsemius

More information about the R-help mailing list