[R] type conversion with apply or not

[William Dunlap]

The short answer is, don't use apply() on data.frame's.
Use lapply to loop over the columns of a data.frame.



> -----Original Message-----
> From: r-help-bounces at r-project.org 
> [mailto:r-help-bounces at r-project.org] On Behalf Of Horace Tso
> Sent: Tuesday, June 08, 2010 2:19 PM
> To: r-help at r-project.org
> Subject: [R] type conversion with apply or not
> 
> Folks, i thought it should be straightforward but after a few 
> hours poking around, I decided it's best to post my question 
> on this list.
> 
> I have a data frame consisting of a (large) number of date 
> columns, which are read in from a csv file as character 
> string. I want to convert them to Date type. Following is an 
> example, where the first column is of integer type, while the 
> rest are type character.
> 
> > head(df)
>   TRANSNO TRANS.START_DATE TRANS.END_DATE DIVISION FASB
> 1  250897         7/1/2010      7/31/2010  PSTRUCT    Z
> 2  250617         8/1/2010      8/31/2010  PSTRUCT    Z
> 3  250364         4/1/2011      6/30/2011      PLR    Z
> 4  250176         4/1/2011      6/30/2011      PLR    Z
> 5  250176         4/1/2011      6/30/2011      PLR    Z
> 6  250364         4/1/2011      6/30/2011      PLR    Z
> > sapply(df, class)
>          TRANSNO TRANS.START_DATE   TRANS.END_DATE         
> DIVISION             FASB
>        "integer"      "character"      "character"      
> "character"      "character"
> I thought it's just a matter of applying with a as.Date,
> 
> df[,c(2,3)] = apply(df[,c(2,3)], 2, function(x)as.Date(x,"%m/%d/%Y"))
> 
> Well, the Date conversion fails and I got,
> 
>   TRANSNO TRANS.START_DATE TRANS.END_DATE DIVISION FASB
> 1  250897            14791          14821  PSTRUCT    Z
> 2  250617            14822          14852  PSTRUCT    Z
> 3  250364            15065          15155      PLR    Z
> 4  250176            15065          15155      PLR    Z
> 5  250176            15065          15155      PLR    Z
> 6  250364            15065          15155      PLR    Z
> The character columns are indeed converted, but they became 
> integer, not Date type. OK, that's strange and so I started 
> reading the help pages.
> 
> It turns out in apply, the result is coerced to some basic 
> vector types.  And apparently Date is not one of the basic 
> vector types.
> 
> "In all cases the result is coerced by 
> as.vector<http://127.0.0.1:19182/library/base/help/as.vector> 
> to one of the basic vector types before the dimensions are 
> set, so that (for example) factor results will be coerced to 
> a character array. "
> 
> The question then is how type conversion can be carried out 
> on some columns of a data frame without using a loop.
> 
> Thanks.
> 
> Horace Tso
> 
> 
> 
> 
> 
> 
> 
> 
> 	[[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.
>