[R] Two time measures

[Gabor Grothendieck]

On Sat, Nov 27, 2010 at 5:02 PM, Eduardo de Oliveira Horta
<eduardo.oliveirahorta at gmail.com> wrote:
> Hello!
>
> I have a csv file of intra-day financial data (5-min closing prices) that
> looks like this: (obs - the dates are formated as day/month/year, as is
> usual here in Brazil)
>
> Date;Time;Close
> 01/09/2009;10:00;56567
> 01/09/2009;10:05;56463
> 01/09/2009;10:10;56370
> ##(goes on all day)
> 01/09/2009;16:45;55771
> 01/09/2009;16:50;55823
> 01/09/2009;16:55;55814
> ##(jumps to the subsequent day)
> 02/09/2009;10:00;55626
> 02/09/2009;10:05;55723
> 02/09/2009;10:10;55659
> ##(goes on all day)
> 02/09/2009;16:45;55742
> 02/09/2009;16:50;55717
> 02/09/2009;16:55;55385
> ## (and so on to the next day)
>

Try this:

library(zoo)
library(chron)
f <- function(x) times(paste(x, 0, sep = ":"))
z <- read.zoo(textConnection(Lines), header = TRUE, sep = ";", split =
1, index = 2, FUN = f)
colnames(z) <- sub("X(..).(..).(....)", "\\3-\\2-\\1", colnames(z))

It produces a zoo object where each column is a series made up of one
day.  The columns are aligned by time of day so the result for the
example data is the following zoo object:

> z
         2009-09-01 2009-09-02
10:00:00      56567      55626
10:05:00      56463      55723
10:10:00      56370      55659
16:45:00      55771      55742
16:50:00      55823      55717
16:55:00      55814      55385




-- 
Statistics & Software Consulting
GKX Group, GKX Associates Inc.
tel: 1-877-GKX-GROUP
email: ggrothendieck at gmail.com