Re: [R] Assigning week numbers
On Wed, Apr 21, 2010 at 6:50 PM, Michael Hosack mhosa...@hotmail.com wrote:
I provided a minimized version of my dataframe at the bottom of this message
containing the results of David's code in variable ('wkoffset') and Jeff
Hallman's code in ('WEEK'). Jeff's code produced the correct results (thank
you Jeff) though I have been unable to understand it. David, as you can see
your code begins week 2 for year 2011 on a Wednesday, rather than on a
Saturday, as it should. Your adjustment seems not to correct the problem, but
I concede I may be using it incorrectly. If you are obtaining the correct
results please let me know what I am doing wrong.
Thanks,
Mike
Hello again,
Just for fun, I implemented the gist of your original code in R. It's
much longer and not as elegant as the other solutions, but perhaps
someone can learn something from it.
Regards,
Gustaf
Daterange-range(SCHEDULE3$DATE.)
Daterange[1]-paste(as.numeric(
substr(as.character(Daterange[1]),1,4))-1,
-05-01,sep=)
Daterange[2]-paste(as.numeric(
substr(as.character(Daterange[2]),1,4))+1,
-05-01,sep=)
alldates-seq(from=Daterange[1],to=Daterange[2],by=1)
My.locale-Sys.getlocale(LC_TIME)
Sys.setlocale(LC_TIME,English_USA.1252)
Week-1
allweeks-vector(length=length(alldates))
for(i in seq_along(alldates)){
if(weekdays(alldates[i])==Saturday){
Week-Week+1
}
if(substr(as.character(alldates[i]),6,10)==05-01){
Week-1
}
allweeks[i]-Week
}
SCHEDULE3$Week-allweeks[match(SCHEDULE3$DATE.,alldates)]
Sys.setlocale(LC_TIME,My.locale)
--
Gustaf Rydevik, M.Sci.
tel: +46(0)703 051 451
address:Essingetorget 40,112 66 Stockholm, SE
skype:gustaf_rydevik
__
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.
Re: [R] Assigning Week Numbers
Take a look at the 'ti' stuff in the tis package on CRAN. If I
understand you correctly, you want something like this:
weekNumber - function(aDate){
aTi - ti(aDate, tif = wfriday)
may1ymd - 1*year(aTi) + 501
baseWeek - ti(may1ymd, tif = wfriday)
return(aTi - baseWeek + 1)
}
--
Jeff
__
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.
Re: [R] Assigning Week Numbers
On Apr 21, 2010, at 9:40 AM, Michael Hosack wrote: Thank you David, this approach is a start in the right direction but it does not yield the needed results. I need new week numbers to only begin on Saturdays. The only exception will be for the first date (May 01) which will start week 1 on a different day of the week depending on the year. The proceeding Saturday will begin Week #2. The approach you provided does not end Week 1 on a Friday and does not end subsequent week #'s on Fridays. My solution should be adjustable for any day-of-week ending number. Why don't you provide a minimal example for testing and show how apply my solution fails? Your original version was definitely not minimal. -- David. Thanks again, Mike CC: mhosa...@hotmail.com; r-help@r-project.org From: dwinsem...@comcast.net To: dwinsem...@comcast.net Subject: Re: [R] Assigning Week Numbers Date: Tue, 20 Apr 2010 16:03:09 -0400 On Apr 20, 2010, at 2:55 PM, David Winsemius wrote: On Apr 20, 2010, at 1:59 PM, Michael Hosack wrote: R experts, How could I extract the week number from a date vector (in Date class) such that week numbering (week 1...2...) begins (May 01) and ends (October 31) on the same specific dates each year? Week numbering must conform to the following day numbering format (Sat=1,Sun=2,Mon=3.Fri=7). This means that new weeks must begin on Saturdays, and end on Fridays (except for the first date of May 01, which always begins week 1; week 2 begins on the proceeding Saturday). This needs to be applicable across years to work effectively. I have tried using both vectorized and loop approaches with no success. Modulo arithmetic will work if you first convert the difftime object to integer: And you can use teh chron function day.of.week to return an offset for the first week weekdays(as.Date(2013-05-01) ) [1] Wednesday day.of.week(2013, 5, 1) [1] 5 SCHEDULE3$wkoffset - as.integer(SCHEDULE3$DATE. - as.Date(2013-05-01)- day.of.week(2013, 5, 1) ) %/% 7 range(SCHEDULE3$wkoffset) [1] -1 25 # So you need a further adjustment of 2 to set the starting week number to 1: SCHEDULE3$wkoffset - 2+as.integer(SCHEDULE3$DATE. - as.Date(2013-05-01)- day.of.week(2013, 5, 1) ) %/% 7 range(SCHEDULE3$wkoffset) [1] 1 27 SCHEDULE3$wkoffset - as.integer(SCHEDULE3$DATE. - as.Date(2010-05-01) ) %/% 7 str(head(SCHEDULE3)) 'data.frame': 6 obs. of 9 variables: $ DATE. :Class 'Date' int [1:6] 15826 15826 15826 15826 15826 15826 $ YEAR : num 2013 2013 2013 2013 2013 ... $ MONTH : num 5 5 5 5 5 5 $ DAY : num 1 1 1 1 1 1 $ DOW : chr Wed Wed Wed Wed ... $ SITE : num 101 101 102 102 103 103 $ TOD : Factor w/ 2 levels MORN,AFTN: 1 2 1 2 1 2 $ DOW. : num 5 5 5 5 5 5 $ wkoffset: num 156 156 156 156 156 156 range(SCHEDULE3$DATE.) [1] 2013-05-01 2013-10-31 range(SCHEDULE3$wkoffset) [1] 156 182 I am including a bit of old Systat code that does the trick simply and concisely. If anyone knows an analogous method in R please let me know. My R dataframe contains all the variables and data in the Systat temp file. Use sched3.t Save sched4.t Hold By mm dd If bof then let week=1 Else if bog and DOW$=SAT then let week = week + 1 Run Thank you, Mike SCHEDULE3 - structure(list(DATE. = structure(c(15826L, 15826L, 15826L, 15826L, snipped the rather larger than minimal example. David Winsemius, MD West Hartford, CT __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Assigning Week Numbers
On Tue, Apr 20, 2010 at 7:59 PM, Michael Hosack mhosa...@hotmail.com wrote: R experts, How could I extract the week number from a date vector (in Date class) such that week numbering (week 1...2...) begins (May 01) and ends (October 31) on the same specific dates each year? Week numbering must conform to the following day numbering format (Sat=1,Sun=2,Mon=3.Fri=7). This means that new weeks must begin on Saturdays, and end on Fridays (except for the first date of May 01, which always begins week 1; week 2 begins on the proceeding Saturday). This needs to be applicable across years to work effectively. I have tried using both vectorized and loop approaches with no success. I am including a bit of old Systat code that does the trick simply and concisely. If anyone knows an analogous method in R please let me know. My R dataframe contains all the variables and data in the Systat temp file. Use sched3.t Save sched4.t Hold By mm dd If bof then let week=1 Else if bog and DOW$=SAT then let week = week + 1 Run Thank you, Mike From your code, it seems as if you're assuming that SCHEDULE3 contains all consecutive saturdays, without skipping any. Is that correct? /Gustaf __ 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] Assigning week numbers
I provided a minimized version of my dataframe at the bottom of this message
containing the results of David's code in variable ('wkoffset') and Jeff
Hallman's code in ('WEEK'). Jeff's code produced the correct results (thank you
Jeff) though I have been unable to understand it. David, as you can see your
code begins week 2 for year 2011 on a Wednesday, rather than on a Saturday, as
it should. Your adjustment seems not to correct the problem, but I concede I
may be using it incorrectly. If you are obtaining the correct results please
let me know what I am doing wrong.
Thanks,
Mike
Jeff Hallman's code:
weeknumber - function(aDate){
aTi - ti(aDate, tif = wfriday)
may1ymd - 1*year(aTi) + 501
baseWeek - ti(may1ymd, tif = wfriday)
return(aTi - baseWeek + 1)
}
SCHEDULE3$WEEK-weeknumber(SCHEDULE3$DATE.)
Thank you David, this approach is a start in the right direction but
it does
not yield the needed results. I need new week numbers to only begin
on Saturdays. The only exception will be for the first date (May 01)
which will start week 1 on a different day of the week depending on
the year. The proceeding Saturday will begin Week #2. The approach you
provided does not end Week 1 on a Friday and does not end subsequent
week #'s on Fridays.
My solution should be adjustable for any day-of-week ending number.
Why don't you provide a minimal example for testing and show how apply
my solution fails? Your original version was definitely not minimal.
--
David.
Thanks again,
Mike
CC: mhosack9 at hotmail.com; r-help at r-project.org
From: dwinsemius at comcast.net
To: dwinsemius at comcast.net
Subject: Re: [R] Assigning Week Numbers
Date: Tue, 20 Apr 2010 16:03:09 -0400
On Apr 20, 2010, at 2:55 PM, David Winsemius wrote:
On Apr 20, 2010, at 1:59 PM, Michael Hosack wrote:
R experts,
How could I extract the week number from a date vector (in Date
class)
such that week numbering (week 1...2...) begins (May 01) and ends
(October 31) on the same specific dates each year? Week numbering
must conform to the following day numbering format
(Sat=1,Sun=2,Mon=3.Fri=7).
This means that new weeks must begin on Saturdays, and end on
Fridays
(except for the first date of May 01, which always begins week 1;
week 2
begins on the proceeding Saturday). This needs to be applicable
across years
to work effectively. I have tried using both vectorized and loop
approaches with
no success.
Modulo arithmetic will work if you first convert the difftime object
to integer:
And you can use teh chron function day.of.week to return an offset
for
the first week
weekdays(as.Date(2013-05-01) )
[1] Wednesday
day.of.week(2013, 5, 1)
[1] 5
SCHEDULE3$wkoffset - as.integer(SCHEDULE3$DATE. -
as.Date(2013-05-01)- day.of.week(2013, 5, 1) ) %/% 7
range(SCHEDULE3$wkoffset)
[1] -1 25
# So you need a further adjustment of 2 to set the starting week
number to 1:
SCHEDULE3$wkoffset - 2+as.integer(SCHEDULE3$DATE. -
as.Date(2013-05-01)- day.of.week(2013, 5, 1) ) %/% 7
range(SCHEDULE3$wkoffset)
[1] 1 27
SCHEDULE3$wkoffset - as.integer(SCHEDULE3$DATE. -
as.Date(2010-05-01) ) %/% 7
str(head(SCHEDULE3))
'data.frame': 6 obs. of 9 variables:
$ DATE. :Class 'Date' int [1:6] 15826 15826 15826 15826 15826 15826
$ YEAR : num 2013 2013 2013 2013 2013 ...
$ MONTH : num 5 5 5 5 5 5
$ DAY : num 1 1 1 1 1 1
$ DOW : chr Wed Wed Wed Wed ...
$ SITE : num 101 101 102 102 103 103
$ TOD : Factor w/ 2 levels MORN,AFTN: 1 2 1 2 1 2
$ DOW. : num 5 5 5 5 5 5
$ wkoffset: num 156 156 156 156 156 156
range(SCHEDULE3$DATE.)
[1] 2013-05-01 2013-10-31
range(SCHEDULE3$wkoffset)
[1] 156 182
I am including a bit of old Systat code that does the trick simply
and concisely.
If anyone knows an analogous method in R please let me know. My R
dataframe contains
all the variables and data in the Systat temp file.
Use sched3.t
Save sched4.t
Hold
By mm dd
If bof then let week=1
Else if bog and DOW$=SAT then let week = week + 1
Run
Thank you,
Mike
SCHEDULE3 -
structure(list(DATE. = structure(c(15826L, 15826L, 15826L, 15826L,
snipped the rather larger than minimal example.
David Winsemius, MD
West Hartford, CT
This is about as minimal as I could get it.
SCHEDULE4 -
structure(list(DATE. = structure(c(15095L, 15095L, 15095L, 15095L,
15095L, 15095L, 15095L, 15095L, 15096L, 15096L, 15096L, 15096L,
15096L, 15096L, 15096L, 15096L, 15097L, 15097L, 15097L, 15097L,
15097L, 15097L, 15097L, 15097L, 15098L, 15098L, 15098L, 15098L,
15098L, 15098L, 15098L, 15098L, 15099L, 15099L, 15099L, 15099L,
15099L, 15099L, 15099L, 15099L, 15100L, 15100L, 15100L, 15100L,
15100L, 15100L, 15100L, 15100L, 15101L, 15101L), class = Date),
YEAR = c(2011, 2011, 2011, 2011, 2011, 2011, 2011, 2011,
2011, 2011, 2011, 2011, 2011, 2011, 2011, 2011, 2011, 2011,
2011, 2011, 2011, 2011, 2011, 2011, 2011, 2011, 2011
Re: [R] Assigning week numbers
On Apr 21, 2010, at 12:50 PM, Michael Hosack wrote:
I provided a minimized version of my dataframe at the bottom of this
message containing the results of David's code in variable
('wkoffset') and Jeff Hallman's code in ('WEEK'). Jeff's code
produced the correct results (thank you Jeff) though I have been
unable to understand it. David, as you can see your code begins week
2 for year 2011 on a Wednesday, rather than on a Saturday, as it
should. Your adjustment seems not to correct the problem, but I
concede I may be using it incorrectly. If you are obtaining the
correct results please let me know what I am doing wrong.
???
require(chron)
SCHEDULE4$wkoffset - 2 + as.integer(SCHEDULE4$DATE. -
as.Date(2011-05-01)- day.of.week(2011, 5, 1) ) %/% 7
tail(SCHEDULE4)
DATE. YEAR MONTH DAY DOW SITE TOD DOW. wkoffset WEEK
374 2011-05-06 2011 5 6 Fri 103 MORN721
1110 2011-05-06 2011 5 6 Fri 103 AFTN721
558 2011-05-06 2011 5 6 Fri 104 MORN721
1294 2011-05-06 2011 5 6 Fri 104 AFTN721
72011-05-07 2011 5 7 Sat 101 MORN122
743 2011-05-07 2011 5 7 Sat 101 AFTN122
I don't know how you are using it, but the registration with a
Saturday==1 seems to be working for me. Perhaps you did not adjust the
year within day.of.week()?
Thanks,
Mike
Jeff Hallman's code:
weeknumber - function(aDate){
aTi - ti(aDate, tif = wfriday)
may1ymd - 1*year(aTi) + 501
baseWeek - ti(may1ymd, tif = wfriday)
return(aTi - baseWeek + 1)
}
SCHEDULE3$WEEK-weeknumber(SCHEDULE3$DATE.)
Thank you David, this approach is a start in the right direction but
it does
not yield the needed results. I need new week numbers to only begin
on Saturdays. The only exception will be for the first date (May 01)
which will start week 1 on a different day of the week depending on
the year. The proceeding Saturday will begin Week #2. The approach
you
provided does not end Week 1 on a Friday and does not end subsequent
week #'s on Fridays.
My solution should be adjustable for any day-of-week ending number.
Why don't you provide a minimal example for testing and show how apply
my solution fails? Your original version was definitely not minimal.
--
David.
Thanks again,
Mike
CC: mhosack9 at hotmail.com; r-help at r-project.org
From: dwinsemius at comcast.net
To: dwinsemius at comcast.net
Subject: Re: [R] Assigning Week Numbers
Date: Tue, 20 Apr 2010 16:03:09 -0400
On Apr 20, 2010, at 2:55 PM, David Winsemius wrote:
On Apr 20, 2010, at 1:59 PM, Michael Hosack wrote:
R experts,
How could I extract the week number from a date vector (in Date
class)
such that week numbering (week 1...2...) begins (May 01) and ends
(October 31) on the same specific dates each year? Week numbering
must conform to the following day numbering format
(Sat=1,Sun=2,Mon=3.Fri=7).
This means that new weeks must begin on Saturdays, and end on
Fridays
(except for the first date of May 01, which always begins week 1;
week 2
begins on the proceeding Saturday). This needs to be applicable
across years
to work effectively. I have tried using both vectorized and loop
approaches with
no success.
Modulo arithmetic will work if you first convert the difftime
object
to integer:
And you can use teh chron function day.of.week to return an offset
for
the first week
weekdays(as.Date(2013-05-01) )
[1] Wednesday
day.of.week(2013, 5, 1)
[1] 5
SCHEDULE3$wkoffset - as.integer(SCHEDULE3$DATE. -
as.Date(2013-05-01)- day.of.week(2013, 5, 1) ) %/% 7
range(SCHEDULE3$wkoffset)
[1] -1 25
# So you need a further adjustment of 2 to set the starting week
number to 1:
SCHEDULE3$wkoffset - 2+as.integer(SCHEDULE3$DATE. -
as.Date(2013-05-01)- day.of.week(2013, 5, 1) ) %/% 7
range(SCHEDULE3$wkoffset)
[1] 1 27
SCHEDULE3$wkoffset - as.integer(SCHEDULE3$DATE. -
as.Date(2010-05-01) ) %/% 7
str(head(SCHEDULE3))
'data.frame': 6 obs. of 9 variables:
$ DATE. :Class 'Date' int [1:6] 15826 15826 15826 15826 15826 15826
$ YEAR : num 2013 2013 2013 2013 2013 ...
$ MONTH : num 5 5 5 5 5 5
$ DAY : num 1 1 1 1 1 1
$ DOW : chr Wed Wed Wed Wed ...
$ SITE : num 101 101 102 102 103 103
$ TOD : Factor w/ 2 levels MORN,AFTN: 1 2 1 2 1 2
$ DOW. : num 5 5 5 5 5 5
$ wkoffset: num 156 156 156 156 156 156
range(SCHEDULE3$DATE.)
[1] 2013-05-01 2013-10-31
range(SCHEDULE3$wkoffset)
[1] 156 182
I am including a bit of old Systat code that does the trick simply
and concisely.
If anyone knows an analogous method in R please let me know. My R
dataframe contains
all the variables and data in the Systat temp file.
Use sched3.t
Save sched4.t
Hold
By mm dd
If bof then let week=1
Else if bog and DOW$=SAT then let week = week + 1
Run
Thank you,
Mike
SCHEDULE3 -
structure(list(DATE. = structure(c(15826L, 15826L, 15826L, 15826L,
snipped the rather larger
[R] Assigning Week Numbers
David, there is still a problem. The data in variable 'wkoffset' should
be equivalent to the data in 'WEEK'. May 07 should begin wkoffset=2 and
the preceeding days should all be assigned wkoffset=1. I did make sure to
adjust the year within day.of.week(). Could you please explain what you
mean by registration with a Saturday==1? Maybe that is where the problem
lies.
Thanks,
Mike
On Apr 21, 2010, at 12:50 PM, Michael Hosack wrote:
I provided a minimized version of my dataframe at the bottom of this
message containing the results of David's code in variable
('wkoffset') and Jeff Hallman's code in ('WEEK'). Jeff's code
produced the correct results (thank you Jeff) though I have been
unable to understand it. David, as you can see your code begins week
2 for year 2011 on a Wednesday, rather than on a Saturday, as it
should. Your adjustment seems not to correct the problem, but I
concede I may be using it incorrectly. If you are obtaining the
correct results please let me know what I am doing wrong.
require(chron)
SCHEDULE4$wkoffset - 2 + as.integer(SCHEDULE4$DATE. -
as.Date(2011-05-01)- day.of.week(2011, 5, 1) ) %/% 7
tail(SCHEDULE4)
DATE. YEAR MONTH DAY DOW SITE TOD DOW. wkoffset WEEK
374 2011-05-06 2011 5 6 Fri 103 MORN721
1110 2011-05-06 2011 5 6 Fri 103 AFTN721
558 2011-05-06 2011 5 6 Fri 104 MORN721
1294 2011-05-06 2011 5 6 Fri 104 AFTN721
72011-05-07 2011 5 7 Sat 101 MORN122
743 2011-05-07 2011 5 7 Sat 101 AFTN122
I don't know how you are using it, but the registration with a
Saturday==1 seems to be working for me. Perhaps you did not adjust the
year within day.of.week()?
Thanks,
Mike
Jeff Hallman's code:
weeknumber - function(aDate){
aTi - ti(aDate, tif = wfriday)
may1ymd - 1*year(aTi) + 501
baseWeek - ti(may1ymd, tif = wfriday)
return(aTi - baseWeek + 1)
}
SCHEDULE3$WEEK-weeknumber(SCHEDULE3$DATE.)
Thank you David, this approach is a start in the right direction but
it does
not yield the needed results. I need new week numbers to only begin
on Saturdays. The only exception will be for the first date (May 01)
which will start week 1 on a different day of the week depending on
the year. The proceeding Saturday will begin Week #2. The approach
you
provided does not end Week 1 on a Friday and does not end subsequent
week #'s on Fridays.
My solution should be adjustable for any day-of-week ending number.
Why don't you provide a minimal example for testing and show how apply
my solution fails? Your original version was definitely not minimal.
--
David.
Thanks again,
Mike
CC: mhosack9 at hotmail.com; r-help at r-project.org
From: dwinsemius at comcast.net
To: dwinsemius at comcast.net
Subject: Re: [R] Assigning Week Numbers
Date: Tue, 20 Apr 2010 16:03:09 -0400
On Apr 20, 2010, at 2:55 PM, David Winsemius wrote:
On Apr 20, 2010, at 1:59 PM, Michael Hosack wrote:
R experts,
How could I extract the week number from a date vector (in Date
class)
such that week numbering (week 1...2...) begins (May 01) and ends
(October 31) on the same specific dates each year? Week numbering
must conform to the following day numbering format
(Sat=1,Sun=2,Mon=3.Fri=7).
This means that new weeks must begin on Saturdays, and end on
Fridays
(except for the first date of May 01, which always begins week 1;
week 2
begins on the proceeding Saturday). This needs to be applicable
across years
to work effectively. I have tried using both vectorized and loop
approaches with
no success.
Modulo arithmetic will work if you first convert the difftime
object
to integer:
And you can use teh chron function day.of.week to return an offset
for
the first week
weekdays(as.Date(2013-05-01) )
[1] Wednesday
day.of.week(2013, 5, 1)
[1] 5
SCHEDULE3$wkoffset - as.integer(SCHEDULE3$DATE. -
as.Date(2013-05-01)- day.of.week(2013, 5, 1) ) %/% 7
range(SCHEDULE3$wkoffset)
[1] -1 25
# So you need a further adjustment of 2 to set the starting week
number to 1:
SCHEDULE3$wkoffset - 2+as.integer(SCHEDULE3$DATE. -
as.Date(2013-05-01)- day.of.week(2013, 5, 1) ) %/% 7
range(SCHEDULE3$wkoffset)
[1] 1 27
SCHEDULE3$wkoffset - as.integer(SCHEDULE3$DATE. -
as.Date(2010-05-01) ) %/% 7
str(head(SCHEDULE3))
'data.frame': 6 obs. of 9 variables:
$ DATE. :Class 'Date' int [1:6] 15826 15826 15826 15826 15826 15826
$ YEAR : num 2013 2013 2013 2013 2013 ...
$ MONTH : num 5 5 5 5 5 5
$ DAY : num 1 1 1 1 1 1
$ DOW : chr Wed Wed Wed Wed ...
$ SITE : num 101 101 102 102 103 103
$ TOD : Factor w/ 2 levels MORN,AFTN: 1 2 1 2 1 2
$ DOW. : num 5 5 5 5 5 5
$ wkoffset: num 156 156 156 156 156 156
range(SCHEDULE3$DATE.)
[1
[R] Assigning week numbers
Gustaf, That is correct. Schedule3 does contain all of the Saturdays between April 30 and Nov. 01 for a given year. Mike R experts, How could I extract the week number from a date vector (in Date class) such that week numbering (week 1...2...) begins (May 01) and ends (October 31) on the same specific dates each year? Week numbering must conform to the following day numbering format (Sat=1,Sun=2,Mon=3.Fri=7). This means that new weeks must begin on Saturdays, and end on Fridays (except for the first date of May 01, which always begins week 1; week 2 begins on the proceeding Saturday). This needs to be applicable across years to work effectively. I have tried using both vectorized and loop approaches with no success. I am including a bit of old Systat code that does the trick simply and concisely. If anyone knows an analogous method in R please let me know. My R dataframe contains all the variables and data in the Systat temp file. Use sched3.t Save sched4.t Hold By mm dd If bof then let week=1 Else if bog and DOW$=SAT then let week = week + 1 Run Thank you, Mike From your code, it seems as if you're assuming that SCHEDULE3 contains all consecutive saturdays, without skipping any. Is that correct? /Gustaf _ The New Busy is not the too busy. Combine all your e-mail accounts with Hotmail. ID28326::T:WLMTAGL:ON:WL:en-US:WM_HMP:042010_4 __ 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.
Re: [R] Assigning week numbers
On Apr 21, 2010, at 1:15 PM, David Winsemius wrote:
On Apr 21, 2010, at 12:50 PM, Michael Hosack wrote:
I provided a minimized version of my dataframe at the bottom of
this message containing the results of David's code in variable
('wkoffset') and Jeff Hallman's code in ('WEEK'). Jeff's code
produced the correct results (thank you Jeff) though I have been
unable to understand it. David, as you can see your code begins
week 2 for year 2011 on a Wednesday, rather than on a Saturday, as
it should. Your adjustment seems not to correct the problem, but I
concede I may be using it incorrectly. If you are obtaining the
correct results please let me know what I am doing wrong.
???
require(chron)
SCHEDULE4$wkoffset - 2 + as.integer(SCHEDULE4$DATE. -
as.Date(2011-05-01)- day.of.week(2011, 5, 1) ) %/% 7
Bad code. I thought the order of arguments to day.of.week would match
the POSIX convention, but it is really month, day, year. Even if I
were on target with the arguments, I would have needed to have a
different adjustment to get the modulo arithmetic working correctly:
dseq - as.Date(2011-05-01) + (-4:14)
dseq
[1] 2011-04-27 2011-04-28 2011-04-29 2011-04-30 2011-05-01
2011-05-02
[7] 2011-05-03 2011-05-04 2011-05-05 2011-05-06 2011-05-07
2011-05-08
[13] 2011-05-09 2011-05-10 2011-05-11 2011-05-12 2011-05-13
2011-05-14
[19] 2011-05-15
data.frame(dseq, wk= 1+ as.integer(dseq - as.Date(2011-05-01)-
day.of.week(5, 1, 2011) +1) %/% 7)
dseq wk
1 2011-04-27 0
2 2011-04-28 0
3 2011-04-29 0
4 2011-04-30 1
5 2011-05-01 1
6 2011-05-02 1
7 2011-05-03 1
8 2011-05-04 1
9 2011-05-05 1
10 2011-05-06 1
11 2011-05-07 2
12 2011-05-08 2
13 2011-05-09 2
14 2011-05-10 2
15 2011-05-11 2
16 2011-05-12 2
17 2011-05-13 2
18 2011-05-14 3
19 2011-05-15 3
So the week counter does now change on a Saturday.
--
David
tail(SCHEDULE4)
DATE. YEAR MONTH DAY DOW SITE TOD DOW. wkoffset WEEK
374 2011-05-06 2011 5 6 Fri 103 MORN721
1110 2011-05-06 2011 5 6 Fri 103 AFTN721
558 2011-05-06 2011 5 6 Fri 104 MORN721
1294 2011-05-06 2011 5 6 Fri 104 AFTN721
72011-05-07 2011 5 7 Sat 101 MORN122
743 2011-05-07 2011 5 7 Sat 101 AFTN122
I don't know how you are using it, but the registration with a
Saturday==1 seems to be working for me.
Only because I was looking a the wrong column. Arrrgghh.
Perhaps you did not adjust the year within day.of.week()?
Thanks,
Mike
Jeff Hallman's code:
weeknumber - function(aDate){
aTi - ti(aDate, tif = wfriday)
may1ymd - 1*year(aTi) + 501
baseWeek - ti(may1ymd, tif = wfriday)
return(aTi - baseWeek + 1)
}
SCHEDULE3$WEEK-weeknumber(SCHEDULE3$DATE.)
Thank you David, this approach is a start in the right direction but
it does
not yield the needed results. I need new week numbers to only begin
on Saturdays. The only exception will be for the first date (May 01)
which will start week 1 on a different day of the week depending on
the year. The proceeding Saturday will begin Week #2. The approach
you
provided does not end Week 1 on a Friday and does not end subsequent
week #'s on Fridays.
My solution should be adjustable for any day-of-week ending number.
Why don't you provide a minimal example for testing and show how
apply
my solution fails? Your original version was definitely not
minimal.
--
David.
Thanks again,
Mike
CC: mhosack9 at hotmail.com; r-help at r-project.org
From: dwinsemius at comcast.net
To: dwinsemius at comcast.net
Subject: Re: [R] Assigning Week Numbers
Date: Tue, 20 Apr 2010 16:03:09 -0400
On Apr 20, 2010, at 2:55 PM, David Winsemius wrote:
On Apr 20, 2010, at 1:59 PM, Michael Hosack wrote:
R experts,
How could I extract the week number from a date vector (in Date
class)
such that week numbering (week 1...2...) begins (May 01) and ends
(October 31) on the same specific dates each year? Week numbering
must conform to the following day numbering format
(Sat=1,Sun=2,Mon=3.Fri=7).
This means that new weeks must begin on Saturdays, and end on
Fridays
(except for the first date of May 01, which always begins week 1;
week 2
begins on the proceeding Saturday). This needs to be applicable
across years
to work effectively. I have tried using both vectorized and loop
approaches with
no success.
Modulo arithmetic will work if you first convert the difftime
object
to integer:
And you can use teh chron function day.of.week to return an offset
for
the first week
weekdays(as.Date(2013-05-01) )
[1] Wednesday
day.of.week(2013, 5, 1)
[1] 5
SCHEDULE3$wkoffset - as.integer(SCHEDULE3$DATE. -
as.Date(2013-05-01)- day.of.week(2013, 5, 1) ) %/% 7
range(SCHEDULE3$wkoffset)
[1] -1 25
# So you need a further adjustment of 2 to set the starting week
number to 1:
SCHEDULE3
