Reading horizontal (row-based) data from xlsx file

Question

This is a "let's try another way" post that is related to this one:

Is it possible to define/modify a reading function that

• can handle the fact that data stored in an xlsx file is row-based (i.e. each row represents a variable)

• and transforms it accordingly so it can be stored in a column-based data.frame (i.e. what used to be a row in xlsx becomes a column)

• while capturing the underlying class/data type of the row-based variables?

Regarding csv files I would probably start with turning to readLines, but unfortunately xlsx is still a black box to me.

Here's a little xlsx file that features examples for both data orientations: https://github.com/rappster/stackoverflow/blob/master/excel/row-and-column-based-data.xlsx

Answer 1

What about slightly modifying the read.xlsx function from the xlsx package:

library(xlsx)
read.transposed.xlsx <- function(file,sheetIndex) {
        df <- read.xlsx(file, sheetIndex = sheetIndex , header = FALSE)
        dft <- as.data.frame(t(df[-1]), stringsAsFactors = FALSE) 
        names(dft) <- df[,1] 
        dft <- as.data.frame(lapply(dft,type.convert))
        return(dft)            
}

# Let's test it
read.transposed.xlsx("row-and-column-based-data.xlsx", sheetIndex = 2)
#    variable var_1 var_2 var_3
#1 2016-01-01     1     a  TRUE
#2 2016-01-02     2     b FALSE
#3 2016-01-03     3     c  TRUE

Answer 2

   > library(openxlsx)
> library(reshape)
> x=read.xlsx("row-and-column-based-data.xlsx",sheet = 2);
> x
  variable 2016-01-01 2016-01-02 2016-01-03
1    var_1          1          2          3
2    var_2          a          b          c
3    var_3       TRUE      FALSE       TRUE
> y=t(x)
> colnames(y)=y[1,]
> y=y[2:nrow(y),]
> cc=data.frame(y, stringsAsFactors = F)
> cc
           var_1 var_2 var_3
2016-01-01     1     a  TRUE
2016-01-02     2     b FALSE
2016-01-03     3     c  TRUE
> sapply(cc, class)
      var_1       var_2       var_3 
"character" "character" "character" 
> write.csv(cc,"temp.csv")
> bb=read.csv("temp.csv")  #infer magically types
> bb
           X var_1 var_2 var_3
1 2016-01-01     1     a  TRUE
2 2016-01-02     2     b FALSE
3 2016-01-03     3     c  TRUE
> sapply(bb, class)
        X     var_1     var_2     var_3 
 "factor" "integer"  "factor" "logical" 

or use stringsAsFactors=F if you prefer character data type:

> bb=read.csv("temp.csv", stringsAsFactors = F)  #infer magically types
> bb
           X var_1 var_2 var_3
1 2016-01-01     1     a  TRUE
2 2016-01-02     2     b FALSE
3 2016-01-03     3     c  TRUE
> sapply(bb, class)
          X       var_1       var_2       var_3 
"character"   "integer" "character"   "logical"