[R] temporal join

[Jeff Ryan]

The data.table package may be more in line with what you are after, but xts
and zoo can also do what you need in this particular example:

> a <- xts(c('a1','a2','a3'), timeBasedSeq(20090101/20090103))
> colnames(a) <- 'foo'

> b <- xts(c('b1'), as.Date('2009-01-04'))
> colnames(b) <- 'foo'
> a
           foo 
2009-01-01 "a1"
2009-01-02 "a2"
2009-01-03 "a3"
> b
           foo 
2009-01-04 "b1"
> cbind(a,b)
           foo  foo.1
2009-01-01 "a1" NA   
2009-01-02 "a2" NA   
2009-01-03 "a3" NA   
2009-01-04 NA   "b1" 
> na.locf(cbind(a,b))['20090104']
           foo  foo.1
2009-01-04 "a3" "b1" 

cbind/merge will merge along the union on the time-index by default (though
all common joins are supported).  The subsetting by time will then find the
dates (or range of dates/times) that match.

na.locf will carry forward last observations.  That is from zoo; which works
on xts, as xts extends zoo.  

HTH,
Jeff




mckenzig wrote:
> 
> I have dataframe a:
> 
> sym date val1
> === ==== ====
> foo 20090101 a1
> foo 20090102 a2
> foo 20090103 a3
> 
> and dataframe b:
> 
> sym date val2
> === ==== ====
> foo 20090104 b1
> 
> I would like to join/merge them to generate the following:
> 
> sym date val2 val1
> === ==== ==== ====
> foo 20090104 b1 a3
> 
> i.e. an equijoin on column 'sym' and a temporal join on column 'date'
> where the closest matching row is retrieved. I have been through the
> various regular/irregular timeseries packages and can not see anything
> like this.
> 
> Regards.
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
> 
> 

-- 
View this message in context: http://www.nabble.com/temporal-join-tp21395414p21437847.html
Sent from the R help mailing list archive at Nabble.com.