[R] Comparing Variables and Writing a New Column

[David Winsemius]

On Feb 1, 2010, at 11:36 AM, David Winsemius wrote:

>
> On Feb 1, 2010, at 11:11 AM, Jerry Floren wrote:
>
>>
>> HI,
>>
>> I am using Windows XP and R version 2.9.2. I have a data frame  
>> written by R
>> similar to the following:
>>
>> Lab_ID    Analysis_Soil        Results        -4MAD      -2.5MAD
>> +2.5MAD        +4MAD
>> 55003    Calcium-2008-116      900             961        1121.5
>> 1656.5          1817
>> 55003    Calcium-2008-117     3321           2175        2380.5
>> 3065.5          3271
>> 55003    Calcium-2008-118     3342           3155        4019
>> 6899            7763
>> 55003    Calcium-2008-119     1664          1005.6      
>> 1147.92       1622.32
>> 1764.64
>> 55003    Calcium-2008-120     2570          1880         
>> 2072            2712
>> 2904
>>
>> Previously, I took this table and finished my analysis in Excel using
>> Excel's "=if" function.
>
> I'll bet that was painful.
>
>> However, I am sure it can be done in R. What I want
>> to do is set up a new data.frame with a new column for Accuracy Flags
>> (a_flag) as shown below.
>>
>> Lab_ID    Analysis_Soil        Results        -4MAD      -2.5MAD
>> +2.5MAD        +4MAD        a_flag
>> 55003    Calcium-2008-116      900             961        1121.5
>> 1656.5          1817         **L
>> 55003    Calcium-2008-117     3321           2175        2380.5
>> 3065.5          3271         **H
>> 55003    Calcium-2008-118     3342           3155        4019
>> 6899            7763           *L
>> 55003    Calcium-2008-119     1664          1005.6      
>> 1147.92       1622.32
>> 1764.64      *H
>> 55003    Calcium-2008-120     2570          1880         
>> 2072            2712
>> 2904
>>
>
> df.soil$a_flag <- apply(df.soil, 1, function(.x)
>          switch(findInterval(.x$Result, c(-Inf,-4,-2.5,2.5,4,Inf)) ,
>                                          "**L", "*L", " ", "*H",  
> "**H") )

Sorry I didn't note the irregularity of the limits. Try instead:

df.soil$a_flag <- apply(df.soil, 1, function(.x)
          switch(findInterval(.x[3], as.numeric( c(-Inf,.x[4:7],Inf)) ),
                                          "**L", "*L", " ", "*H",  
"**H") )

I wasn't sure why the passage of the function made it necessary to add  
the as.numeric, but I am guessing that .x became a character vector in  
order to hold the first two columns. That would not be be needed, it  
you only passed the last 5 columns as df.soil[3:7] as an argument to  
apply.
>
>
>
>> For each row I need to compare the "Results" submitted by the labs  
>> to the
>> four "MAD" columns. If the Results are less than -4.0 MAD units  
>> from the
>> median, labs are flagged "**L" (very low). For results greater than  
>> +4.0 MAD
>> units, labs are flagged "**H" (very high). Likewise for -2.5 MAD  
>> and +2.5
>> MAD (*L and *H respectively). As shown in the last row, labs are  
>> not flagged
>> for results within -2.5 MAD to +2.5 MAD units.
>>
>> Can anyone get me started on how to look at each row and compare the
>> "Results" variable with each of the four "__MAD" variables and then  
>> writing
>> the appriate flag for Results exceeding -2.5 MAD to +2.5 MAD units  
>> from the
>> median?
>>
>> Thanks,
>>
>> Jerry Floren
>> Minnesota Department of Agriculture
>>
>>
>>
>>
>>
>>
>>
>> -- 
>> View this message in context: http://n4.nabble.com/Comparing-Variables-and-Writing-a-New-Column-tp1458947p1458947.html
>> Sent from the R help mailing list archive at Nabble.com.
>>
>> ______________________________________________
>> 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
>
> ______________________________________________
> 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