How to convert a 3 dimensional List to make a table with tables tabular()? - R-help

Beckstrand, Janis, NCOD · 2013-11-18T20:46:09Z

An embedded and charset-unspecified text was scrubbed... Name: not available URL:

Tue, Nov 19, 2013 3:27 AM #

On 13-11-18 3:46 PM, Beckstrand, Janis, NCOD wrote:

I have information in a 3 dimensional list that I want to use to create
a table for a written report.



The 3d list was created by applying the epiR function " epitest " to a
list of 8 2x2 tables using  lapply.

The resulting list gives the 12 statistics with CIs,  that  generated by
epitest, for each of the 8 matrices (i9, i10, etc).



The list has the following form:



Statslow$i9

$aprev

         est     lower    upper

1 0.2263884 0.2244714 0.228314



$tprev

        est     lower     upper

1 0.204388 0.2025413 0.2062442



Statslow$i10

$aprev

         est     lower    upper

1 0.227004 0.22545  0.228314



$tprev

        est     lower     upper

1 0.204388 0.2025413 0.2062442





I want to create a report that has a table in the following format:

                                    I9
i10

                  Est   lower     upper                   Est  lower
upper



aprev   0.23     0.22     0.23                           etc

tprev    0.2       0.20     0.21

etc



I have read and tried the tables package pdf and also the information at

http://www.r-statistics.com/tag/data-frame/ about cast() and
tabular.cast_df.r(),

and also all the information I can find on nested tables, but have not
be successful.

I have tried to use melt() on the 3D list (to create a dataframe
"statslow.d")  but get what seems more like S3 output :



   variable     value           L2                L1

1      est       0.2263884    aprev       i9

2    lower    0.2244714    aprev      i9

3    upper   0.2283140    aprev       i9

4      est       0.2043880     tprev      i9

5    lower    0.2025413    tprev       i9

6    upper   0.2062442    tprev       i9



When I try to use it with tables() I get errors

and I can't see how to use the tables function on it to get what I want.

Here is the tables() code that I have tried:



class(statslow.d)

[1] "data.frame"

head(statslow.d)

   variable     value    L2  L1

1      est 0.2263884 aprev  i9

2    lower 0.2244714 aprev i9

3    upper 0.2283140 aprev i9

4      est 0.2043880 tprev i9

5    lower 0.2025413 tprev i9

6    upper 0.2062442 tprev i9

tabular(statslow.d,L2~L1*variable)

Error in tabular.formula(as.formula(table, env = parent.frame()), ...) :


   data must be a dataframe, list or environment

statslow.d<-as.data.frame(statslow.d)

tabular(statslow.d,L2~L1*variable)

Error in tabular.formula(as.formula(table, env = parent.frame()), ...) :


   data must be a dataframe, list or environment



I would appreciate any help or hints anyone can give.

The arguments to tabular are in the wrong order.  You need to give the 
formula first or name it as "table".  So

tabular(L2 ~ L1*variable, data=statslow.d)

would not give an error, but it probably doesn't do what you want, 
because you don't specify a statistic, so it will default to a count. 
You could use the "mean" function, e.g.

tabular(L2 ~ L1*variable*Heading()*mean, data=statslow.d)

(The Heading() will suppress the column label "mean".)  There are 
examples in the vignette that do fancy formatting of multiple columns, 
so it would be possible to write the interval as [.224, .228] for instance.

Duncan Murdoch