Factors are you friend here:
myData
mydate gender mygroup id mygrp.f
1 2012-03-25 F A 1 1
2 2005-05-23 F B 2 2
3 2005-09-08 F B 2 2
4 2005-12-07 F B 2 2
5 2006-02-26 F C 2 3
6 2006-05-13 F
I have a data set with about 6 million rows and 50 columns. It is a
mixture of dates, factors, and numerics.
What I am trying to accomplish can be seen with the following
simplified data, which is given as dput output below.
head(myData)
mydate gender mygroup id
1 2012-03-25 F
Hi Juliet:
Here's a Q D solution:
# (1) plyr
f - function(d) length(unique(d$mygroup)) - 1
ddply(myData, .(id), f)
id V1
1 1 0
2 2 2
3 3 1
4 4 0
# (2) data.table
myDT - data.table(myData, key = 'id')
myDT[, list(nswitch = length(unique(mygroup)) - 1), by = 'id']
If one can switch
Thanks Dennis! I'll check this out.
Just to clarify, I need the total number of switches/changes
regardless of if that state
had occurred in the past. So A-A-B-A, would have 2 changes: A to B and B to A.
Thanks again.
On Wed, Aug 24, 2011 at 1:28 PM, Dennis Murphy djmu...@gmail.com wrote:
Hi
4 matches
Mail list logo
