[R] 'matrix' returns integer instead of decimal

[Marc Schwartz]

Brant Inman wrote:
> R-helpers:
> 
>  I am experiencing some odd behavior with the 'matrix' function that I have
> not experienced before and was wondering if there is something that I was
> missing in my code.
> 
>  ---------------------------------
>  > sessionInfo()
> R version 2.6.1 (2007-11-26)
> i386-pc-mingw32
>  locale:
> LC_COLLATE=English_United States.1252;LC_CTYPE=English_United
> States.1252;LC_MONETARY=English_United
> States.1252;LC_NUMERIC=C;LC_TIME=English_United States.1252
>  attached base packages:
> [1] splines   grid      stats     graphics  grDevices utils     datasets
>  methods
> [9] base
>  other attached packages:
> [1] meta_0.8-2         car_1.2-7          arm_1.1-1          R2WinBUGS_2.1-6
> 
> [5] coda_0.13-1        lme4_0.99875-9     Matrix_0.999375-3  nlme_3.1-86
> 
> [9] latticeExtra_0.3-1 RColorBrewer_1.0-2 survival_2.34      MASS_7.2-39
> 
> [13] foreign_0.8-23     RGraphics_1.0-6    lattice_0.17-4
>  loaded via a namespace (and not attached):
> [1] rcompgen_0.1-17 tools_2.6.1
> 
>  > rm(list=ls())
> 
>> matrix(c(7,47,4,38,20,96,1,14,10,48,2,101,12,161,     # Make a data matrix
> + 1,28,1,19,22,49,25,162,31,200,9,39,22,193,
> + 0.5,45.5,31,131,4,75,31,220,7,55,3,91,14.5,
> + 25.5,3,65,25,54,24,41,37,95,11,17,26,49,13,
> + 84,38,170,29,60,9,20,44,47,30,160,40,185,10,
> + 41,40,185,4.5,46.5,60,140,12,75,42,225,26,57,
> + 17,92,23.5,23.5,6,68), ncol=2, byrow=T)
> 
>      [,1] [,2]
> [1,]  7.0   47
> [2,]  4.0   38
> [3,] 20.0   96
> [4,]  1.0   14
> [5,] 10.0   48
> [6,]  2.0  101
> [7,] 12.0  161
> [8,]  1.0   28
> [9,]  1.0   19
> [10,] 22.0   49
> [11,] 25.0  162
> [12,] 31.0  200
> [13,]  9.0   39
> [14,] 22.0  193
> [15,]  0.5   46
> [16,] 31.0  131
> [17,]  4.0   75
> [18,] 31.0  220
> [19,]  7.0   55
> [20,]  3.0   91
> [21,] 14.5   26
> [22,]  3.0   65
> [23,] 25.0   54
> [24,] 24.0   41
> [25,] 37.0   95
> [26,] 11.0   17
> [27,] 26.0   49
> [28,] 13.0   84
> [29,] 38.0  170
> [30,] 29.0   60
> [31,]  9.0   20
> [32,] 44.0   47
> [33,] 30.0  160
> [34,] 40.0  185
> [35,] 10.0   41
> [36,] 40.0  185
> [37,]  4.5   46
> [38,] 60.0  140
> [39,] 12.0   75
> [40,] 42.0  225
> [41,] 26.0   57
> [42,] 17.0   92
> [43,] 23.5   24
> [44,]  6.0   68
> 
> ------------------------------------
> 
> Note the problems in rows 15, 21, 37 and 43.  They should read [0.5, 45.5],
> [14.5, 21.5], etc...  The matrix function seems to be rounding the second
> column up to the next integer.  Why would this occur?  Can I do something to
> prevent this?
> 
> I would very much appreciate any comments.
> 
> Thanks,
> 
> Brant Inman
> Mayo Clinic

Hi Brant,

I don't get this with 2.6.1 patched on Fedora 8:

mat <- matrix(c(7,47,4,38,20,96,1,14,10,48,2,101,12,161,
                1,28,1,19,22,49,25,162,31,200,9,39,22,193,
                0.5,45.5,31,131,4,75,31,220,7,55,3,91,14.5,
                25.5,3,65,25,54,24,41,37,95,11,17,26,49,13,
                84,38,170,29,60,9,20,44,47,30,160,40,185,10,
                41,40,185,4.5,46.5,60,140,12,75,42,225,26,57,
                17,92,23.5,23.5,6,68), ncol=2, byrow=T)

> mat
      [,1]  [,2]
 [1,]  7.0  47.0
 [2,]  4.0  38.0
 [3,] 20.0  96.0
 [4,]  1.0  14.0
 [5,] 10.0  48.0
 [6,]  2.0 101.0
 [7,] 12.0 161.0
 [8,]  1.0  28.0
 [9,]  1.0  19.0
[10,] 22.0  49.0
[11,] 25.0 162.0
[12,] 31.0 200.0
[13,]  9.0  39.0
[14,] 22.0 193.0
[15,]  0.5  45.5
[16,] 31.0 131.0
[17,]  4.0  75.0
[18,] 31.0 220.0
[19,]  7.0  55.0
[20,]  3.0  91.0
[21,] 14.5  25.5
[22,]  3.0  65.0
[23,] 25.0  54.0
[24,] 24.0  41.0
[25,] 37.0  95.0
[26,] 11.0  17.0
[27,] 26.0  49.0
[28,] 13.0  84.0
[29,] 38.0 170.0
[30,] 29.0  60.0
[31,]  9.0  20.0
[32,] 44.0  47.0
[33,] 30.0 160.0
[34,] 40.0 185.0
[35,] 10.0  41.0
[36,] 40.0 185.0
[37,]  4.5  46.5
[38,] 60.0 140.0
[39,] 12.0  75.0
[40,] 42.0 225.0
[41,] 26.0  57.0
[42,] 17.0  92.0
[43,] 23.5  23.5
[44,]  6.0  68.0

Given some of the packages you have listed above, I would try to go into
R using:

  R --vanilla

from a command line and then try your example again. It seems possible
that one of the packages has modified the behavior of print() on a
matrix or is causing some type of conflict.

For example, if options("digits") has been modified from it's default of
'7':

> options("digits")
$digits
[1] 7

to either 1 or 2, I can replicate the behavior you see above:

options(digits = 2)

> mat
      [,1] [,2]
 [1,]  7.0   47
 [2,]  4.0   38
 [3,] 20.0   96
 [4,]  1.0   14
 [5,] 10.0   48
 [6,]  2.0  101
 [7,] 12.0  161
 [8,]  1.0   28
 [9,]  1.0   19
[10,] 22.0   49
[11,] 25.0  162
[12,] 31.0  200
[13,]  9.0   39
[14,] 22.0  193
[15,]  0.5   46
[16,] 31.0  131
[17,]  4.0   75
[18,] 31.0  220
[19,]  7.0   55
[20,]  3.0   91
[21,] 14.5   26
[22,]  3.0   65
[23,] 25.0   54
[24,] 24.0   41
[25,] 37.0   95
[26,] 11.0   17
[27,] 26.0   49
[28,] 13.0   84
[29,] 38.0  170
[30,] 29.0   60
[31,]  9.0   20
[32,] 44.0   47
[33,] 30.0  160
[34,] 40.0  185
[35,] 10.0   41
[36,] 40.0  185
[37,]  4.5   46
[38,] 60.0  140
[39,] 12.0   75
[40,] 42.0  225
[41,] 26.0   57
[42,] 17.0   92
[43,] 23.5   24
[44,]  6.0   68



HTH,

Marc Schwartz