Re: [R] get start and end date of ISO weeks giving a date as input
Hello Enrico,
2016-09-08 10:41 GMT-03:00 Enrico Schumann :
> Hi Veronica,
>
> please see inline.
>
> On Thu, 08 Sep 2016, Veronica Andreo writes:
>
> > Hello Luisfo and Enrico,
> >
> > Thanks for your help! I've been testing both
> > solutions... results differ for the same date (I
> > changed both functions to use ISO8601). And I added
> > contiguous dates, to see how they handle the
> > start-end of the week.
> >
> > So, here the results:
> >
> > ### one example
> > d <- c("2010-08-21","2010-08-22","2010-08-23","2010-08-24")
> > iso_start_end <- function(d) {
> > d <- as.Date(d)
> > wday <- as.POSIXlt(d)$wday
> > data.frame(date = d,
> > week = format(d, "%V"),
> > starts = d - wday + 1,
> > ends = d + 7 - wday)
> > }
> > iso_start_end(d)
> >
> > date week starts ends
> > 1 2010-08-21 33 2010-08-16 2010-08-22
> > 2 2010-08-22 33 2010-08-23 2010-08-29
> > 3 2010-08-23 34 2010-08-23 2010-08-29
> > 4 2010-08-24 34 2010-08-23 2010-08-29
>
> Yes, the second date makes no sense, and it happens
> because Sunday is 0 (and not 7). My bad. Here is
> a fixed version:
>
> iso_start_end <- function(d) {
> d <- as.Date(d)
> wday <- as.POSIXlt(d)$wday
> wday[wday == 0] <- 7
> data.frame(date = d,
> week = format(d, "%V"),
> starts = d - wday + 1,
> ends = d + 7 - wday)
> }
>
>
> > ### the other example:
> > dd <- as.Date(strptime('2010-08-21', format="%Y-%m-%d", tz="GMT"))
> > ref.date <- as.Date(strptime(paste0(year(dd),"-01-01"),
> format="%Y-%m-%d"))
> > bound.dates <- ref.date + 7 * (isoweek(dd)) + c(0,6)
> > bound.dates
> > [1] "2010-08-20" "2010-08-26"
>
> You can use the function "weekdays" to see check the
> results.
>
> > weekdays(bound.dates)
> [1] "Friday" "Thursday"
>
> > So, researching a bit more and inspired by those
> > examples, I eventually came up with this solution
> > that seems to work fine... I share in case that any
> > other has a similar problem:
> >
> > # get ISOweek for my vector of dates
> > week_iso<-ISOweek(d)
> >
> > # vector with the format %Y-W%V-1 for start day of the ISO week
> > week_iso_day1 <- paste(week_iso,1, sep="-")
> >
> > # vector with the format %Y-W%V-7 for end day of the ISO week
> > week_iso_day7 <- paste(week_iso, 7, sep="-")
> >
> > # use ISOweek2date
> > data.frame(date= d, week_iso = week_iso, start =
> ISOweek2date(week_iso_day1), end = ISOweek2date(week_iso_day7)
> >
> > date week_iso startend
> > 1 2010-08-21 2010-W33 2010-08-16 2010-08-22
> > 2 2010-08-22 2010-W33 2010-08-16 2010-08-22
> > 3 2010-08-23 2010-W34 2010-08-23 2010-08-29
> > 4 2010-08-24 2010-W34 2010-08-23 2010-08-29
>
> The updated 'iso_start_end' gives the same result.
>
> date week starts ends
> 1 2010-08-21 33 2010-08-16 2010-08-22
> 2 2010-08-22 33 2010-08-16 2010-08-22
> 3 2010-08-23 34 2010-08-23 2010-08-29
> 4 2010-08-24 34 2010-08-23 2010-08-29
>
Yes! Again, thanks for your time and help!
Best,
Vero
[[alternative HTML version deleted]]
__
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
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.
Re: [R] get start and end date of ISO weeks giving a date as input
Hello Luisfo and Enrico,
Thanks for your help! I've been testing both solutions... results differ
for the same date (I changed both functions to use ISO8601). And I added
contiguous dates, to see how they handle the start-end of the week.
So, here the results:
### one example
d <- c("2010-08-21","2010-08-22","2010-08-23","2010-08-24")
iso_start_end <- function(d) {
d <- as.Date(d)
wday <- as.POSIXlt(d)$wday
data.frame(date = d,
week = format(d, "%V"),
starts = d - wday + 1,
ends = d + 7 - wday)
}
iso_start_end(d)
date week starts ends
1 2010-08-21 33 2010-08-16 2010-08-22
*2 2010-08-22 33 2010-08-23 2010-08-29*
3 2010-08-23 34 2010-08-23 2010-08-29
4 2010-08-24 34 2010-08-23 2010-08-29
### the other example:
dd <- as.Date(strptime('2010-08-21', format="%Y-%m-%d", tz="GMT"))
ref.date <- as.Date(strptime(paste0(year(dd),"-01-01"), format="%Y-%m-%d"))
bound.dates <- ref.date + 7 * (isoweek(dd)) + c(0,6)
bound.dates
[1] "2010-08-20" "2010-08-26"
So, researching a bit more and inspired by those examples, I eventually
came up with this solution that seems to work fine... I share in case that
any other has a similar problem:
# get ISOweek for my vector of dates
week_iso<-ISOweek(d)
# vector with the format %Y-W%V-1 for start day of the ISO week
week_iso_day1 <- paste(week_iso,1, sep="-")
# vector with the format %Y-W%V-7 for end day of the ISO week
week_iso_day7 <- paste(week_iso, 7, sep="-")
# use ISOweek2date
data.frame(date= d, week_iso = week_iso, start =
ISOweek2date(week_iso_day1), end = ISOweek2date(week_iso_day7)
date week_iso startend
1 2010-08-21 2010-W33 2010-08-16 2010-08-22
2 2010-08-22 2010-W33 2010-08-16 2010-08-22
3 2010-08-23 2010-W34 2010-08-23 2010-08-29
4 2010-08-24 2010-W34 2010-08-23 2010-08-29
Thanks again for your time, ideas and help!
Best,
Vero
2016-09-08 8:20 GMT-03:00 Luisfo :
> Dear Veronica,
>
> Here there's a way of doing what you requested.
>
> library("lubridate")
> # your date '2010-08-21' as Date object
> dd <- as.Date(strptime("2010-08-21", format="%Y-%m-%d", tz="GMT"))
> # take the first day of the year as Date object, i.e. 2010-01-01 in our
> example
> ref.date <- as.Date(strptime(paste0(year(dd),"-01-01"),
> format="%Y-%m-%d", tz="GMT"))
> # the start and end dates
> bound.dates <- ref.date + 7 * (week(dd)-1) + c(0,6)
>
> I hope you find it useful.
>
> Best,
> *Luisfo Chiroque*
>
> *PhD Student | PhD Candidate IMDEA Networks Institute*
> http://fourier.networks.imdea.org/people/~luis_nunez/
>
> On 09/08/2016 12:13 PM, Veronica Andreo wrote:
>
> Hello list,
>
> Is there a quick way to get start and end date (%Y-%m-%d) from ISO
> weeks if I only have dates?
>
> For example, I have this date in which some event happened:
> "2010-08-21". Not only I want the ISO week, which I can obtain either
> with isoweek (lubridate) or ISOweek (ISOweek), but I want the start
> and end date of that ISO week.
>
> Do I need to print all ISO weeks from the period of interest and
> sample there for start and end date? Or is there a better way to do
> that?
>
> Thanks a lot in advance!
>
> Best,
> Veronica
>
> __r-h...@r-project.org mailing
> list -- To UNSUBSCRIBE and more,
> seehttps://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.
>
>
>
[[alternative HTML version deleted]]
__
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
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.
[R] get start and end date of ISO weeks giving a date as input
Hello list, Is there a quick way to get start and end date (%Y-%m-%d) from ISO weeks if I only have dates? For example, I have this date in which some event happened: "2010-08-21". Not only I want the ISO week, which I can obtain either with isoweek (lubridate) or ISOweek (ISOweek), but I want the start and end date of that ISO week. Do I need to print all ISO weeks from the period of interest and sample there for start and end date? Or is there a better way to do that? Thanks a lot in advance! Best, Veronica __ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see 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.
[R] multiple model selection for vglm models
Hi list, I'm working with vglm (family tobit) models and I need to perform multiple model selection. Something similar to step or stepAIC, is there anything alike? I found that neither step nor stepAIC work for objects of class vglm. Has anybody worked with them and can give a hint on how to select among vglm models? Thanks a lot in advance, Vero [[alternative HTML version deleted]] __ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see 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.
Re: [R] help with script to get starting date of blooms
Hi Jim
2014-07-19 8:25 GMT-03:00 Jim Lemon :
>
> Hi Vero,
> I think this does what you want, even though it doesn't do it the way
> you describe above.
>
> cla<-runif(506)+
> rep(c(seq(1,3,length.out=23),seq(3,1,length.out=23)),11)
> obs_time<-paste(rep(1:11,each=46),rep(1:46,11),sep="_")
> t_max<-rep(0,506)
> t_begin<-rep("",11)
> for(year in 1:11) {
> threshold<-median(cla[(year-1) * 46 + 1:46]) * 1.05
> max_pos<-(year-1) * 46 + which.max(cla[(year-1)*46+1:46])
> t_max[max_pos]<-1
> threshpos<-(year-1) * 46 + 1
> while(cla[threshpos] < threshold) threshpos <- threshpos + 1
> cat(threshold,cla[threshpos],max_pos,cla[max_pos],"\n")
> t_begin[year]<-obs_time[threshpos]
> }
>
Yes, indeed!!! It does the job beautifully!!!
Thanks so much!
I'll now try to adapt it to the 150.000 time series that form my study area
in South Atlantic Ocean :)
Any advice?
Best,
Vero
[[alternative HTML version deleted]]
__
R-help@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.
[R] help with script to get starting date of blooms
Hi list
I have a vector, which are remotely sensed chlorophyll values for a certain
pixel in 11 years, then i have a flag or label vector (t_max) that consists
of 0 and 1, which tells me where i have the annual peak (one per year, then
eleven 1 and all the rest are 0).
I'm interested in extracting the date in which the bloom beggins. For that
matter I'm using a threshold of 5% above the median of the whole series.
I need to go over the chlorophyll data vector (cla) and every time I find
the yearly maximum value (where t_max == 1), i need to go backwards in cla
as many time steps as needed untill I cross the threshold (5% over the
median of cl)... once i crossed that threshold, i need to be sure that at
least for 2 time steps, the value of cla keeps going down (it's below the
threshold)... once that condition is met, i want to get the first position
where cla is higher than that threshold...
that would be the phytoplankton bloom starting date...
This is more or less what i was trying, but it does not work and i'm
stuck... can you please help me??
cla<- ts of 506 values (every 46 values, i have one year)
t_max<- vector of 0 & 1; 1 where the yearly cla max occurs (11 ones & all
rest 0)
threshold = 0.05*median(cla)+median(cla)
for ( t in 1:length(t_max) ) {
if ( t_max[t] == 0 )
{
next;
} else {
if ( (cla[t-1] < threshold) && (cla[t-2)] < threshold) )
{
return which(cla[t])
}
}
}
Thanks a lot in advance!! :)
Vero
[[alternative HTML version deleted]]
__
R-help@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.
