[R] Replace missing value within group with non-missing value
Leask, Graham
g.leask at aston.ac.uk
Sat Apr 6 19:23:26 CEST 2013
Hi David,
Thank you for your suggestion unfortunately it yields an incorrect result.
Within a choice set the mth should be identical.
> head(dat)
dn obs choice br mth
1 4 1 0 1 487
2 4 1 0 2 488
3 4 1 0 3 488
4 4 1 0 4 488
5 4 1 0 5 488
6 4 1 1 6 487
Best wishes
Graham
-----Original Message-----
From: David Winsemius [mailto:dwinsemius at comcast.net]
Sent: 06 April 2013 17:46
To: David Winsemius
Cc: Leask, Graham; r-help at r-project.org
Subject: Re: [R] Replace missing value within group with non-missing value
On Apr 6, 2013, at 9:26 AM, David Winsemius wrote:
>
> On Apr 6, 2013, at 9:16 AM, Leask, Graham wrote:
>
>> Hi Rui,
>>
>> Data as follows
>>
>> structure(list(dn = c(4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
>> 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
>> 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4), obs = c(1, 1, 1, 1, 1, 1, 2, 2, 2,
>> 2, 2, 2, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 6, 6,
>> 6, 6, 6, 6, 7, 7, 7, 7, 7, 7, 8, 8, 8, 8, 8, 8, 9, 9), choice = c(0,
>> 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0,
>> 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,
>> 1, 0, 0), br = c(1, 2, 3, 4, 5, 6, 1, 2, 3, 4, 5, 6, 1, 2, 3, 4, 5,
>> 6, 1, 2, 3, 4, 5, 6, 1, 2, 3, 4, 5, 6, 1, 2, 3, 4, 5, 6, 1, 2, 3, 4,
>> 5, 6, 1, 2, 3, 4, 5, 6, 1, 2), mth = c(NA, NA, NA, NA, NA, 487, NA,
>> NA, 488, NA, NA, NA, NA, NA, NA, NA, NA, 488, NA, NA, 489, NA, NA,
>> NA, NA, NA, NA, NA, NA, 489, NA, NA, NA, NA, NA, 489, NA, NA, NA, NA,
>> NA, 490, NA, NA, NA, NA, NA, 491, NA, NA)), .Names = c("dn", "obs",
>> "choice", "br", "mth"), row.names = c("1", "2", "3", "4", "5", "6",
>> "7", "8", "9", "10", "11", "12", "13", "14", "15", "16", "17", "18",
>> "19", "20", "21", "22", "23", "24", "25", "26", "27", "28", "29",
>> "30", "31", "32", "33", "34", "35", "36", "37", "38", "39", "40",
>> "41", "42", "43", "44", "45", "46", "47", "48", "49", "50"), class =
>> "data.frame")
>>
>
> Looks like a job for na.locf in the zoo package:
>
> require(zoo)
> # will fail if first value is NA so either this ...
> dat$mth[-(1:5)] <- na.locf(dat$mth[-(1:5)]) # ... or this:
> dat$mth <- na.locf(dat$mth, na.rm=FALSE)
If on the other hand you wnated cases to be be handled within individual values of "obs" then you could do this to the categories of obs where there was a value to replace (omitting the last two rows):
dat$mth[-(49:50)] <- ave(dat$mth[-(49:50)] , dat$obs,
FUN=function(m) {m[is.na(m)] <- m[!is.na(m)]; m } )
If there were more than one choice in a category you might need to pick the first or the last.
>
> --
> David.
>
>> Best wishes
>>
>>
>> Graham
>>
>> -----Original Message-----
>> From: Rui Barradas [mailto:ruipbarradas at sapo.pt]
>> Sent: 06 April 2013 16:32
>> To: Leask, Graham
>> Cc: r-help at r-project.org
>> Subject: Re: [R] Replace missing value within group with non-missing
>> value
>>
>> Hello,
>>
>> Can't you post a data example? If your dataset is named 'dat' use
>>
>> dput(head(dat, 50)) # paste the output of this in a post
>>
>>
>> Rui Barradas
>>
>> Em 06-04-2013 15:34, Leask, Graham escreveu:
>>> Hi Rui,
>>>
>>> Thank you for your suggestion which is very much appreciated. Unfortunately running this code produces the following error.
>>>
>>> error in '$<-.data.frame' ('*tmp*', "mth", value = NA_real_) :
>>> replacement has 1 rows, data has 0
>>>
>>> I'm sure there must be an elegant solution to this problem?
>>>
>>> Best wishes
>>>
>>>
>>>
>>> Graham
>>>
>>> On 6 Apr 2013, at 12:15, "Rui Barradas" <ruipbarradas at sapo.pt> wrote:
>>>
>>>> Hello,
>>>>
>>>> That's not a very good way of posting your data, preferably paste the output of ?dput in a post.
>>>> Some thing along the lines of the following might do what you want.
>>>> It seems that the groups are established by 'dn' and 'obs' numbers.
>>>> If so, try
>>>>
>>>>
>>>> # Make up some data
>>>> dat <- data.frame(dn = 4, obs = rep(1:5, each = 6), mth = NA)
>>>> dat$mth[6] <- 487 dat$mth[9] <- 488 dat$mth[18] <- 488 dat$mth[21]
>>>> <-
>>>> 489 dat$mth[30] <- 489
>>>>
>>>>
>>>> sp <- split(dat, list(dat$dn, dat$obs))
>>>> names(sp) <- NULL
>>>> tmp <- lapply(sp, function(x){
>>>> idx <- which(!is.na(x$mth))[1]
>>>> x$mth <- x$mth[idx]
>>>> x
>>>> })
>>>> do.call(rbind, tmp)
>>>>
>>>>
>>>> Hope this helps,
>>>>
>>>> Rui Barradas
>>>>
>>>>
>>>> Em 06-04-2013 11:33, Leask, Graham escreveu:
>>>>> Dear List members
>>>>>
>>>>> I have a large dataset organised in choice groups see sample below
>>>>>
>>>>> +-------------------------------------------------------------------------------------------------+
>>>>> | dn obs choice acid br date cdate situat~n mth year set |
>>>>> |-------------------------------------------------------------------------------------------------|
>>>>> 1. | 4 1 0 LOSEC 1 . . . . 1 |
>>>>> 2. | 4 1 0 NEXIUM 2 . . . . 1 |
>>>>> 3. | 4 1 0 PARIET 3 . . . . 1 |
>>>>> 4. | 4 1 0 PROTIUM 4 . . . . 1 |
>>>>> 5. | 4 1 0 ZANTAC 5 . . . . 1 |
>>>>> |-------------------------------------------------------------------------------------------------|
>>>>> 6. | 4 1 1 ZOTON 6 23aug2000 01:00:00 23aug2000 NS 487 2000 1 |
>>>>> 7. | 4 2 0 LOSEC 1 . . . . 2 |
>>>>> 8. | 4 2 0 NEXIUM 2 . . . . 2 |
>>>>> 9. | 4 2 1 PARIET 3 25sep2000 01:00:00 25sep2000 L 488 2000 2 |
>>>>> 10. | 4 2 0 PROTIUM 4 . . . . 2 |
>>>>> |-------------------------------------------------------------------------------------------------|
>>>>> 11. | 4 2 0 ZANTAC 5 . . . . 2 |
>>>>> 12. | 4 2 0 ZOTON 6 . . . . 2 |
>>>>> 13. | 4 3 0 LOSEC 1 . . . . 3 |
>>>>> 14. | 4 3 0 NEXIUM 2 . . . . 3 |
>>>>> 15. | 4 3 0 PARIET 3 . . . . 3 |
>>>>> |-------------------------------------------------------------------------------------------------|
>>>>> 16. | 4 3 0 PROTIUM 4 . . . . 3 |
>>>>> 17. | 4 3 0 ZANTAC 5 . . . . 3 |
>>>>> 18. | 4 3 1 ZOTON 6 20sep2000 00:00:00 20sep2000 R 488 2000 3 |
>>>>> 19. | 4 4 0 LOSEC 1 . . . . 4 |
>>>>> 20. | 4 4 0 NEXIUM 2 . . . . 4 |
>>>>> |-------------------------------------------------------------------------------------------------|
>>>>> 21. | 4 4 1 PARIET 3 27oct2000 00:00:00 27oct2000 NL 489 2000 4 |
>>>>> 22. | 4 4 0 PROTIUM 4 . . . . 4 |
>>>>> 23. | 4 4 0 ZANTAC 5 . . . . 4 |
>>>>> 24. | 4 4 0 ZOTON 6 . . . . 4 |
>>>>> 25. | 4 5 0 LOSEC 1 . . . . 5 |
>>>>> |-------------------------------------------------------------------------------------------------|
>>>>> 26. | 4 5 0 NEXIUM 2 . . . . 5 |
>>>>> 27. | 4 5 0 PARIET 3 . . . . 5 |
>>>>> 28. | 4 5 0 PROTIUM 4 . . . . 5 |
>>>>> 29. | 4 5 0 ZANTAC 5 . . . . 5 |
>>>>> 30. | 4 5 1 ZOTON 6 23oct2000 03:00:00 23oct2000 NS 489 2000 5 |
>>>>>
>>>>> I wish to fill in the missing values in each choice set - delineated by dn (Doctor) obs (Observation number) and choices (1 to 6).
>>>>> For each choice set one choice is chosen which contains full time
>>>>> information for that choice set ie in set 1 choice 6 was chosen and shows the month 487. The other 5 choices show mth as missing. I want to fill these with the correct mth.
>>>>>
>>>>> I am sure there must be an elegant way to do this in R?
>>>>>
>>>>>
>>>>> Best wishes
>>>>>
>>>>>
>>>>>
>>>>> Graham
>>>>>
>>>>>
>>>>> [[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.
>>>>
>>>
>>
>> ______________________________________________
>> 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
> Alameda, CA, USA
>
> ______________________________________________
> 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
Alameda, CA, USA
More information about the R-help
mailing list