[R] Syntax Check: rshape2 melt()

[Nordlund, Dan (DSHS/RDA)]

> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-
> project.org] On Behalf Of Rich Shepard
> Sent: Thursday, October 27, 2011 2:41 PM
> To: r-help at r-project.org
> Subject: Re: [R] Syntax Check: rshape2 melt()
> 
> On Thu, 27 Oct 2011, Dennis Murphy wrote:
> 
> > # Define param so that all of its levels are represented:
> > tds.a <- within(tds.a, {
> >           param = factor(param, levels = c('TDS', 'Cond', 'Mg', 'Ca',
> > 'Cl', 'Na', 'SO4'))
> >           sampdate = as.Date(sampdate)  } )
> 
> Dennis,
> 
>    This produces an error:
> tds.a <- within(tds.anal, {param = factor(param, levels = c('TDS',
> 'Cond',
> 'Mg', 'Ca', 'Cl', 'Na', 'SO4')) sampdate = as.Date(sampdate)})
> Error: unexpected symbol in "tds.a <- within(tds.anal, {param =
> factor(param, levels = c('TDS', 'Cond', 'Mg', 'Ca', 'Cl', 'Na', 'SO4'))
> sampdate"
> 
>    So I did the conversion and melting the long way:
> > tds.a <- tds.anal
> > tds.a$sampdate <- factor(tds.a$sampdate)
> 
>    Now sampdate is a factor, and quant has the proper values:
> str(tds.a)
> 'data.frame':	11434 obs. of  4 variables:
>   $ site    : Factor w/ 143 levels "BC-0.5","BC-1",..: 134 127 134 134
> 73 73
>   $ sampdate: Factor w/ 890 levels "1981-11-30","1982-04-28",..: 686
> 268 727
>   $ param   : Factor w/ 7 levels "Ca","Cl","Cond",..: 7 3 4 6 1 2 4 5 1
> 4 ...
>   $ quant   : num  10800 280 1620 7580 79 114 301 27.5 381 274 ...
> 
>   However, ...
> 
> > dcast(tds.a, site + sampdate ~ param, value_var = 'quant')
> 
> c.tds.anal <- dcast(tds.a, site + sampdate ~ param, value_var =
> 'quant')
> Aggregation function missing: defaulting to length.
> 
>    That's because c.tds.anal has lost the quant values:
> 
> str(c.tds.anal)
> 'data.frame':	2309 obs. of  9 variables:
>   $ site    : Factor w/ 143 levels "BC-0.5","BC-1",..: 1 1 1 2 2 2 2 2
> 2 2
> ...
>   $ sampdate: Factor w/ 890 levels "1981-11-30","1982-04-28",..: 239
> 252 260
> 99 101 102 155 163 170 179 ...
>   $ Ca      : int  1 0 0 1 1 1 1 1 0 1 ...
>   $ Cl      : int  1 0 0 1 1 1 1 1 0 1 ...
>   $ Cond    : int  1 1 1 1 1 1 1 1 1 1 ...
>   $ Mg      : int  1 0 0 1 1 1 1 1 0 1 ...
>   $ Na      : int  0 0 0 1 1 1 1 1 0 1 ...
>   $ SO4     : int  1 1 1 1 1 1 1 1 0 1 ...
>   $ TDS     : int  1 1 1 1 1 1 1 1 0 1 ...
> 
>    I don't see what I did differently from your sample code.
> 
> Rich
> 

Rich,

It appears that what you did differently was to dcast the un-melted data.  In addition, you shouldn't need to turn your dates into factors.  I assume that your unaltered data is in tds.anal.

Melt the data this way:

>m.tds.anal <-  melt(tds.anal, id.vars = c('site', 'sampdate', 'param'))

Then, cast the data this way:

>wanted <- dcast(m.tds.anal, site + sampdate ~ param)


Does this do what you wanted?


Dan

Daniel J. Nordlund
Washington State Department of Social and Health Services
Planning, Performance, and Accountability
Research and Data Analysis Division
Olympia, WA 98504-5204