R help - Jun 2009 - Merging Irregular Time Series With NAs

Hi

I have two irregular time series, which are of different lengths and being  
and end at different times. For the common subset of time that they both  
span, they should have the same values, but the values may occur at  
slightly different time intervals. I am trying to "line up" the
identical
values and reconcile them. I have merged the two series into a zoo object  
which looks like the following:
> head(m)
aq$mid d2$mid
2009-06-22 16:25:40.044 NA 1.63755
2009-06-22 16:25:40.909 1.63760 NA
2009-06-22 16:25:40.987 NA 1.63760
2009-06-22 16:25:41.657 1.63755 NA
2009-06-22 16:25:41.738 NA 1.63755
2009-06-22 16:25:41.909 1.63760 NA

What I would like to do is merge the series column-wise : ie where one  
column contains NAs, replace the NAs with the non-NA value from the other  
column. Is this possible easily?

Here is a dput() of the data above:

structure(c(NA, 1.6376, NA, 1.63755, NA, 1.6376, 1.63755, NA,
1.6376, NA, 1.63755, NA), .Dim = c(6L, 2L), .Dimnames = list(
NULL, c("aq$mid", "d2$mid")), index =
structure(c(1245684340.044,
1245684340.909, 1245684340.987, 1245684341.657, 1245684341.738,
1245684341.909), class = c("POSIXt", "POSIXct")), class =
"zoo")


Cheers
-- Rory

	[[alternative HTML version deleted]]

Your zoo object is illegal since zoo objects are not supposed to
have duplicated times; however, ignoring that if the object has
at most one NA in each row, which is the case in the example,
then this will produce a numeric vector and you can create a
new zoo object out of that:
> rowSums(z, na.rm = TRUE)
[1] 1.63755 1.63760 1.63760 1.63755 1.63755 1.63760


On Tue, Jun 23, 2009 at 9:29 AM, <rory.winston at gmail.com>
wrote:
> Hi
>
> I have two irregular time series, which are of different lengths and being
> and end at different times. For the common subset of time that they both
> span, they should have the same values, but the values may occur at
> slightly different time intervals. I am trying to "line up" the
identical
> values and reconcile them. I have merged the two series into a zoo object
> which looks like the following:
>
>> head(m)
> aq$mid d2$mid
> 2009-06-22 16:25:40.044 NA 1.63755
> 2009-06-22 16:25:40.909 1.63760 NA
> 2009-06-22 16:25:40.987 NA 1.63760
> 2009-06-22 16:25:41.657 1.63755 NA
> 2009-06-22 16:25:41.738 NA 1.63755
> 2009-06-22 16:25:41.909 1.63760 NA
>
> What I would like to do is merge the series column-wise : ie where one
> column contains NAs, replace the NAs with the non-NA value from the other
> column. Is this possible easily?
>
> Here is a dput() of the data above:
>
> structure(c(NA, 1.6376, NA, 1.63755, NA, 1.6376, 1.63755, NA,
> 1.6376, NA, 1.63755, NA), .Dim = c(6L, 2L), .Dimnames = list(
> NULL, c("aq$mid", "d2$mid")), index =
structure(c(1245684340.044,
> 1245684340.909, 1245684340.987, 1245684341.657, 1245684341.738,
> 1245684341.909), class = c("POSIXt", "POSIXct")), class
= "zoo")
>
>
> Cheers
> -- Rory
>
> ? ? ? ?[[alternative HTML version deleted]]
>
> ______________________________________________
> 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.
>

On Jun 23, 2009, at 9:29 AM, rory.winston at gmail.com wrote:
> Hi
>
> I have two irregular time series, which are of different lengths and  
> being
> and end at different times. For the common subset of time that they  
> both
> span, they should have the same values, but the values may occur at
> slightly different time intervals. I am trying to "line up" the  
> identical
> values and reconcile them. I have merged the two series into a zoo  
> object
> which looks like the following:
>
>> head(m)
> aq$mid d2$mid
> 2009-06-22 16:25:40.044 NA 1.63755
> 2009-06-22 16:25:40.909 1.63760 NA
> 2009-06-22 16:25:40.987 NA 1.63760
> 2009-06-22 16:25:41.657 1.63755 NA
> 2009-06-22 16:25:41.738 NA 1.63755
> 2009-06-22 16:25:41.909 1.63760 NA
>
> What I would like to do is merge the series column-wise : ie where one
> column contains NAs, replace the NAs with the non-NA value from the  
> other
> column. Is this possible easily?
Depends what you define as easy:

require(zoo)
  m[is.na(m[,1]),1] <- m[is.na(m[,1]),2]
  m[is.na(m[,2]),2] <- m[is.na(m[,2]),1]

 > m
                      aq$mid  d2$mid
2009-06-22 11:25:40 1.63755 1.63755
2009-06-22 11:25:40 1.63760 1.63760
2009-06-22 11:25:40 1.63760 1.63760
2009-06-22 11:25:41 1.63755 1.63755
2009-06-22 11:25:41 1.63755 1.63755
2009-06-22 11:25:41 1.63760 1.63760
>
> Here is a dput() of the data above:
>
> structure(c(NA, 1.6376, NA, 1.63755, NA, 1.6376, 1.63755, NA,
> 1.6376, NA, 1.63755, NA), .Dim = c(6L, 2L), .Dimnames = list(
> NULL, c("aq$mid", "d2$mid")), index =
structure(c(1245684340.044,
> 1245684340.909, 1245684340.987, 1245684341.657, 1245684341.738,
> 1245684341.909), class = c("POSIXt", "POSIXct")), class
= "zoo")
>
>
> Cheers
> -- Rory
>
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> 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.
David Winsemius, MD
Heritage Laboratories
West Hartford, CT

On Jun 23, 2009, at 9:50 AM, Gabor Grothendieck wrote:
> Your zoo object is illegal since zoo objects are not supposed to
> have duplicated times;
They were unique at the sub-second level.
> however, ignoring that if the object has
> at most one NA in each row, which is the case in the example,
> then this will produce a numeric vector and you can create a
> new zoo object out of that:
>
>> rowSums(z, na.rm = TRUE)
> [1] 1.63755 1.63760 1.63760 1.63755 1.63755 1.63760
But that will not give the desired answer if both aq$mid and d2$mid  
were present at the same time.
>
>
> On Tue, Jun 23, 2009 at 9:29 AM, <rory.winston at gmail.com> wrote:
>> Hi
>>
>> I have two irregular time series, which are of different lengths  
>> and being
>> and end at different times. For the common subset of time that they  
>> both
>> span, they should have the same values, but the values may occur at
>> slightly different time intervals. I am trying to "line up"
the
>> identical
>> values and reconcile them. I have merged the two series into a zoo  
>> object
>> which looks like the following:
>>
>>> head(m)
>> aq$mid d2$mid
>> 2009-06-22 16:25:40.044 NA 1.63755
>> 2009-06-22 16:25:40.909 1.63760 NA
>> 2009-06-22 16:25:40.987 NA 1.63760
>> 2009-06-22 16:25:41.657 1.63755 NA
>> 2009-06-22 16:25:41.738 NA 1.63755
>> 2009-06-22 16:25:41.909 1.63760 NA
>>
>> What I would like to do is merge the series column-wise : ie where  
>> one
>> column contains NAs, replace the NAs with the non-NA value from the  
>> other
>> column. Is this possible easily?
>>
>> Here is a dput() of the data above:
>>
>> structure(c(NA, 1.6376, NA, 1.63755, NA, 1.6376, 1.63755, NA,
>> 1.6376, NA, 1.63755, NA), .Dim = c(6L, 2L), .Dimnames = list(
>> NULL, c("aq$mid", "d2$mid")), index =
structure(c(1245684340.044,
>> 1245684340.909, 1245684340.987, 1245684341.657, 1245684341.738,
>> 1245684341.909), class = c("POSIXt", "POSIXct")),
class = "zoo")
>>
>>
>> Cheers
>> -- Rory
>>
David Winsemius, MD
Heritage Laboratories
West Hartford, CT

On Tue, Jun 23, 2009 at 10:04 AM, David Winsemius<dwinsemius at
comcast.net> wrote:
>
> On Jun 23, 2009, at 9:50 AM, Gabor Grothendieck wrote:
>
>> Your zoo object is illegal since zoo objects are not supposed to
>> have duplicated times;
>
> They were unique at the sub-second level.
>
>> however, ignoring that if the object has
>> at most one NA in each row, which is the case in the example,
>> then this will produce a numeric vector and you can create a
>> new zoo object out of that:
>>
>>> rowSums(z, na.rm = TRUE)
>>
>> [1] 1.63755 1.63760 1.63760 1.63755 1.63755 1.63760
>
> But that will not give the desired answer if both aq$mid and d2$mid were
> present at the same time.
What was written was:
"if the object has at most one NA in each row, which is the case in
the example, ..."

On Tue, Jun 23, 2009 at 10:10 AM, Gabor
Grothendieck<ggrothendieck at gmail.com> wrote:
> On Tue, Jun 23, 2009 at 10:04 AM, David Winsemius<dwinsemius at
comcast.net> wrote:
>>
>> On Jun 23, 2009, at 9:50 AM, Gabor Grothendieck wrote:
>>
>>> Your zoo object is illegal since zoo objects are not supposed to
>>> have duplicated times;
>>
>> They were unique at the sub-second level.
>>
>>> however, ignoring that if the object has
>>> at most one NA in each row, which is the case in the example,
>>> then this will produce a numeric vector and you can create a
>>> new zoo object out of that:
>>>
>>>> rowSums(z, na.rm = TRUE)
>>>
>>> [1] 1.63755 1.63760 1.63760 1.63755 1.63755 1.63760
>>
>> But that will not give the desired answer if both aq$mid and d2$mid
were
>> present at the same time.
>
> What was written was:
> "if the object has at most one NA in each row, which is the case in
> the example, ..."
>
hit send too quickly.  continuing:

but what was meant was at most one non-NA.