Zero column data frames make addDataFrame fail

[GoogleCodeExporter]

What steps will reproduce the problem?
wb <- createWorkbook()
sheet <- createSheet(wb)
dfr <- data.frame(x = 1:5)[,FALSE] #Data frame with some rows, no columns.
addDataFrame(dfr, sheet)
## Error in .jcall(cellBlock$ref, "V", setDataMethod, as.integer(j - 1L),  : 
##   java.lang.ArrayIndexOutOfBoundsException: 1

What is the expected output? What do you see instead?
I expected addDataFrame to gracefully decline to add any content to the 
worksheet rather than throwing an error.

What version of the product are you using? On what operating system?
xlsx version '0.5.5'.

Please provide any additional information below.

The problem is caused by line 46 (depending upon how you count them):

    for (j in 1:ncol(x)) {

When x has zero columns, then 1:ncol(x) is c(1, 0), not an empty vector.

To fix the problem, change the line to:

    for(j in seq_along(x)) {

Original issue reported on code.google.com by richiero...@gmail.com on 30 Jan 2014 at 11:32

[GoogleCodeExporter]

A quick search through the source of xlsx reveals that the file write.xlsx.R 
contains two more instances of 1:ncol, and five instances of 1:nrow.  I haven't 
checked to see if these cause problems, but it is safer to replaces them with 
calls to seq_len or seq_along.

Original comment by richiero...@gmail.com on 30 Jan 2014 at 11:35

[GoogleCodeExporter]

thanks.  looking at it.

Original comment by adrian.d...@gmail.com on 13 Feb 2014 at 4:12

• Changed state: Accepted

[GoogleCodeExporter]

Fixed in trunk. 

Original comment by adrian.d...@gmail.com on 13 Feb 2014 at 6:43

• Changed state: Done