[R] Replace NAs in split lists

[Ek Esawi]

Thank you Jeff. Your code works, as usual , perfectly. I am just
wondering why if i put the whole code in one line, i get an error
message.
sdf2 <- lapply( sdf, function(z){z$Value
<-ifelse(is.na(z$Value),z$Value[!is.na(z$Value)][1],z$Value)z})
error. unexpected symbol in sdf2

Thanks again

EK


On Mon, Jan 8, 2018 at 3:12 AM, Jeff Newmiller <jdnewmil at dcn.davis.ca.us> wrote:
> Upon closer examination I see that you are not using the split version of
> df1 as I usually would, so here is a reproducible example:
>
> #----
> df1 <- read.table( text=
> "ID ID_2 Firist Value
> 1  a   aa   TRUE     2
> 2  a   ab  FALSE    NA
> 3  a   ac  FALSE    NA
> 4  b   aa   TRUE     5
> 5  b   ab  FALSE    NA
> ", header=TRUE, as.is=TRUE )
>
> sdf <- split( df1, df1$ID )
> # note the extra [ 1 ] in case you have more than one non-NA value # per ID
> sdf2 <- lapply( sdf
>               , function( z ) {
>                  z$Value <- ifelse( is.na( z$Value )
>                                   , z$Value[ !is.na( z$Value ) ][ 1 ]
>                                   , z$Value
>                                   )
>                  z
>                 }
>               )
> df2 <- do.call( rbind, sdf2 )
> df2
> #>     ID ID_2 Firist Value
> #> a.1  a   aa   TRUE     2
> #> a.2  a   ab  FALSE     2
> #> a.3  a   ac  FALSE     2
> #> b.4  b   aa   TRUE     5
> #> b.5  b   ab  FALSE     5
>
> # or using tidyverse methods
>
> library(dplyr)
> #>
> #> Attaching package: 'dplyr'
> #> The following objects are masked from 'package:stats':
> #>
> #>     filter, lag
> #> The following objects are masked from 'package:base':
> #>
> #>     intersect, setdiff, setequal, union
> df3 <- (   df1
>        %>% group_by( ID )
>        %>% do({
>               mutate( .
>                     , Value = ifelse( is.na( Value )
>                                     , Value[ !is.na( Value ) ][ 1 ]
>                                     , Value
>                                     )
>                     )
>            })
>        %>% ungroup
>        )
> df3
> #> # A tibble: 5 x 4
> #>   ID    ID_2  Firist Value
> #>   <chr> <chr> <lgl>  <int>
> #> 1 a     aa    T          2
> #> 2 a     ab    F          2
> #> 3 a     ac    F          2
> #> 4 b     aa    T          5
> #> 5 b     ab    F          5
> #----
>
>
> On Sun, 7 Jan 2018, Jeff Newmiller wrote:
>
>> Why do you want to modify df1?
>>
>> Why not just reassemble the parts as a new data frame and use that going
>> forward in your calculations? That is generally the preferred approach in R
>> so you can re-do your calculations easily if you find a mistake later.
>> --
>> Sent from my phone. Please excuse my brevity.
>>
>> On January 7, 2018 7:35:59 PM PST, Ek Esawi <esawiek at gmail.com> wrote:
>>>
>>> I just came up with a solution right after i posted the question, but
>>> i figured there must be a better and shorter one.than my solution
>>> sdf1[[1]][1,4]<-lapplyresults[[1]]
>>> sdf1[[2]][1,4]<-lapplyresults[[2]]
>>>
>>> EK
>>>
>>> On Sun, Jan 7, 2018 at 10:13 PM, Ek Esawi <esawiek at gmail.com> wrote:
>>>>
>>>> Hi all--
>>>>
>>>> I stumbled on this problem online. I did not like the solution given
>>>> there which was a long UDF. I thought why cannot split and l/s apply
>>>> work here. My aim is to split the data frame, use l/sapply, make
>>>> changes on the split lists and combine the split lists to new data
>>>> frame with the desired changes/output.
>>>>
>>>> The data frame shown below has a column named ID which has 2
>>>
>>> variables
>>>>
>>>> a and b; i want to replace the NAs on the Value column by 2, which is
>>>> the only numeric entry, for ID=a and by 5 for ID=b.
>>>>
>>>> I worked out the solution but could not replace the results in the
>>>
>>> split lists.
>>>>
>>>>
>>>> Original dataframe , df1
>>>>   ID ID_2 Firist Value
>>>> 1  a   aa   TRUE     2
>>>> 2  a   ab  FALSE    NA
>>>> 3  a   ac  FALSE    NA
>>>> 4  b   aa   TRUE     5
>>>> 5  b   ab  FALSE    NA
>>>> Sdf1
>>>> $a
>>>> ID ID_2 Firist Value
>>>> 1  a   aa   TRUE     2
>>>> 2  a   ab  FALSE    NA
>>>> 3  a   ac  FALSE    NA
>>>> $b
>>>>   ID ID_2 Firist Value
>>>> 4  b   aa   TRUE     5
>>>> 5  b   ab  FALSE    NA
>>>> Desired results
>>>> ID ID_2 Firist Value
>>>> 1  a   aa   TRUE    2
>>>> 2  a   ab  FALSE    2
>>>> 3  a   ac  FALSE    2
>>>>
>>>> $b
>>>>   ID ID_2 Firist Value
>>>> 4  b   aa   TRUE     5
>>>> 5  b   ab  FALSE     5
>>>>
>>>> My code
>>>>
>>>> sdf <- split(df1,df$ID)
>>>> lapply(sdf, function(z)
>>>
>>> ifelse(is.na(z$Value),z$Value[!is.na(z$Value)],z$Value))
>>>>
>>>> result:
>>>> $ a: num [1:3] 2 2 2
>>>> $ b: num [1:2] 5 5
>>>>
>>>> How could I put these two lists back in the split data frame, sdf1?
>>>> Then I could use do.call to reassemble a data frame from the split
>>>> lists,
>>>>
>>>> Thanks,
>>>> EK
>>>
>>>
>>> ______________________________________________
>>> R-help at 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-help at 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.
>>
>
> ---------------------------------------------------------------------------
> Jeff Newmiller                        The     .....       .....  Go Live...
> DCN:<jdnewmil at dcn.davis.ca.us>        Basics: ##.#.       ##.#.  Live Go...
>                                       Live:   OO#.. Dead: OO#..  Playing
> Research Engineer (Solar/Batteries            O.O#.       #.O#.  with
> /Software/Embedded Controllers)               .OO#.       .OO#.  rocks...1k
> ---------------------------------------------------------------------------