[R] NA problem when use paste function

[Prof Brian Ripley]

On Wed, 16 Apr 2008, Lu, Jiang wrote:

> Dear R helpers,
>
> I was doing a genetic project with two datasets X and Y. There are
> some IDs in both data sets, and others in either data set. I used
> "merge(x,y,by="ID",all=TRUE)". The data set Y contains a variable (a
> genotype) which is also in data X. When I merge X with Y, these two
> variables were automatically re-named by appending .x and .y to the
> original variable names. As you can see on the following list, I would
> like to take whatever available (non-missing non-NA) in X or Y as the
> final value for the genotype S3Allel1. I used paste() function.
> However, it converts <NA> to NA as character. Would you please tell me
> how I can just get the genotype without pasting the NA to it? I
> checked the document of paste() and noticed that it used
> as.character() to the vector argument. I guess that is the reason I
> got "NA" as a string for the new variable I created (S3Allele1).

Please don't 'guess': that is not what as.character does.

Your example is not reproducible (see the footer of this message) and it 
is not clear what the structure is.  But <NA> indicates a missing value in 
a factor or unquoted character vector.  E.g.

> x <- c("G", "A", "A")
> y <- rep(NA_character_, 3)
> data.frame(x, y)
   x    y
1 G <NA>
2 A <NA>
3 A <NA>
> paste(x, y)
[1] "G NA" "A NA" "A NA"

Here y does contain missing values and paste() converted them to "NA".
As the help says:

      Note that 'paste()' coerces 'NA_character_', the character missing
      value, to '"NA"' which may seem undesirable, e.g., when pasting
      two character vectors, or very desirable, e.g. in 'paste("the
      value of p is ", p)'.

Possibly you want

ifelse(is.na(x), y, x)



> Should I use any other funtion to avoid this problem? Any insight is
> appreciated!
>
>           ID      S3Allele1.x S3Allele1.y S3Allele1
> 1       10003           G        <NA>      G NA
> 2       10004           A        <NA>      A NA
> 3       10005           A        <NA>      A NA
> 4       10006           A        <NA>      A NA
> 5       10007           G        <NA>      G NA
> 6       10008           A        <NA>      A NA
> 7       10009           A        <NA>      A NA
> 8       10010           A        <NA>      A NA
> 9       10011           A        <NA>      A NA
> 10      10013           A        <NA>      A NA
> 11      10014           A        <NA>      A NA
> 12      10015           A        <NA>      A NA
> 13      10016           A        <NA>      A NA
> 14      10017           A        <NA>      A NA
> 15      10018           A        <NA>      A NA
> 16      10019           G        <NA>      G NA
> 17      10020           A        <NA>      A NA
> 18      10021           G        <NA>      G NA
> 19      10022           A        <NA>      A NA
> 20      10023           G        <NA>      G NA
> 21      10024           G        <NA>      G NA
> 22      10025           G        <NA>      G NA
> 23      10027           G        <NA>      G NA
> 24      10028           G        <NA>      G NA
> 25      10029           G        <NA>      G NA
> 26      10031           G        <NA>      G NA
> 27      10032           A        <NA>      A NA
> 28      10033        <NA>                   NA
> 29      10035           A        <NA>      A NA
> 30      10037           A        <NA>      A NA
> 31      10038        <NA>           A      NA A
> 32      10039        <NA>           A      NA A
>
> ______________________________________________
> 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.
>

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595