[R] Split a dataframe by rownames and/or colnames

[David Winsemius]

On Feb 20, 2015, at 9:33 AM, Tim Richter-Heitmann wrote:

> Dear List,
> 
> Consider this example
> 
> df <- data.frame(matrix(rnorm(9*9), ncol=9))
> names(df) <- c("c_1", "d_1", "e_1", "a_p", "b_p", "c_p", "1_o1", "2_o1", "3_o1")
> row.names(df) <- names(df)
> 
> 
> indx <- gsub(".*_", "", names(df))
> 
> I can split the dataframe by the index that is given in the column.names after the underscore "_".
> 
> list2env(
>  setNames(
>    lapply(split(colnames(df), indx), function(x) df[x]),
>    paste('df', sort(unique(indx)), sep="_")),
>  envir=.GlobalEnv)
> 

> However, i changed my mind and want to do it now by rownames. Exchanging colnames with rownames does not work, it gives the exact same output (9 rows x 3 columns). I could do
> as.data.frame(t(df_x),
> but maybe that is not elegant.
> What would be the solution for splitting the dataframe by rows?

The split.data.frame method seems to work perfectly well with a rownames-derived index argument:

> split(df, sub(".+_","", rownames(df) ) )
$`1`
      c_1   d_1  e_1   a_p   b_p   c_p  1_o1 2_o1  3_o1
c_1 -0.11 -0.04 1.33 -0.87 -0.16 -0.25 -0.75 0.34  0.14
d_1 -0.62 -0.94 0.80 -0.78 -0.70  0.74  0.11 1.44 -0.33
e_1  0.98 -0.83 0.48  0.19 -0.32 -1.01  1.28 1.04 -2.16

$o1
       c_1   d_1   e_1   a_p   b_p   c_p  1_o1  2_o1  3_o1
1_o1 -0.93 -0.02  0.69 -0.67  1.04  1.04 -1.50 -0.36  0.50
2_o1  0.02 -0.16 -0.09 -1.50 -0.02 -1.04  1.07 -0.45  1.56
3_o1 -1.42  0.88 -0.05  0.85 -1.35  0.21  1.35  0.92 -0.76

$p
      c_1   d_1   e_1   a_p  b_p   c_p  1_o1  2_o1  3_o1
a_p -1.35  0.91 -0.58 -0.63 0.94 -1.13  0.71  0.25  0.82
b_p -0.25 -0.73 -0.41 -1.71 1.28  0.19 -0.35  1.74 -0.93
c_p -0.01 -1.11 -0.12  0.58 1.51  0.03 -0.99 -0.23 -0.03

> 
> Thank you very much!
> 
> -- 
> Tim Richter-Heitmann
> 
-- 

David Winsemius
Alameda, CA, USA