R help - Sep 2016 - `head` doesn't show all columns for an empty data.frame

I'm sure I'm doing something wrong, but I'm seeing strange behaviour
using the `head` and `tail` functions on an empty data.frame.

To reproduce:
# create an empty data frame.  I actually read an empty table from Excel using
`readWorkbook` from package `openxlsx`
test <- structure(list(Code = NULL, Name = NULL, Address = NULL, Sun.Hrs =
NULL,
    Mon.Hrs = NULL), .Names = c("Code", "Name",
"Address", "Sun.Hrs",
"Mon.Hrs"), class = "data.frame", row.names = integer(0))

# show the data frame
test

# output in console:
#     [1] Code    Name    Address Sun.Hrs Mon.Hrs
#     <0 rows> (or 0-length row.names)

# note that the data frame has 0 rows and 5 columns
# show the structure
str(test)

# output in console:
#'data.frame':	0 obs. of  5 variables:
# $ Code   : NULL
# $ Name   : NULL
# $ Address: NULL
# $ Sun.Hrs: NULL
# $ Mon.Hrs: NULL

#again, the structure shows 5 columns.  However...
head(test); tail(test)

# output in console:
#[1] Name    Sun.Hrs
#<0 rows> (or 0-length row.names)
#[1] Name    Sun.Hrs
#<0 rows> (or 0-length row.names)

# now we have only two columns
 
Weird, right?

So, here's my session info:
> sessionInfo()
R version 3.3.1 (2016-06-21)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows >= 8 x64 (build 9200)

locale:
[1] LC_COLLATE=English_United States.1252  LC_CTYPE=English_United States.1252  
LC_MONETARY=English_United States.1252 LC_NUMERIC=C
[5] LC_TIME=English_United States.1252    

attached base packages:
[1] stats4    grid      stats     graphics  grDevices utils     datasets 
methods   base

other attached packages:
 [1] tidyr_0.6.0       lpSolve_5.6.13    flexclust_1.3-4   modeltools_0.2-21
lattice_0.20-34   gtools_3.5.0      reshape2_1.4.1    ash_1.0-15       
RODBC_1.3-13
[10] ggmap_2.6.1       ggplot2_2.1.0     dplyr_0.5.0       assertthat_0.1   
openxlsx_3.0.0

loaded via a namespace (and not attached):
 [1] Rcpp_0.12.7       plyr_1.8.4        tools_3.3.1       digest_0.6.10    
tibble_1.2        gtable_0.2.0      png_0.1-7         DBI_0.5-1        
mapproj_1.2-4
[10] parallel_3.3.1    proto_0.3-10      stringr_1.1.0     RgoogleMaps_1.4.1
maps_3.1.1        R6_2.1.3          jpeg_0.1-8        sp_1.2-3         
magrittr_1.5
[19] scales_0.4.0      geosphere_1.5-5   colorspace_1.2-6  labeling_0.3     
stringi_1.1.1     lazyeval_0.2.0    munsell_0.4.3     rjson_0.2.15

This is not an urgent issue, I just think it's curious, so it would be nice
to understand why it happens.

Thanks,

Colin

On 22/09/2016 4:45 AM, Colin Phillips wrote:
> I'm sure I'm doing something wrong, but I'm seeing strange
behaviour using the `head` and `tail` functions on an empty data.frame.
>
> To reproduce:
> # create an empty data frame.  I actually read an empty table from Excel
using `readWorkbook` from package `openxlsx`
> test <- structure(list(Code = NULL, Name = NULL, Address = NULL, Sun.Hrs
= NULL,
>     Mon.Hrs = NULL), .Names = c("Code", "Name",
"Address", "Sun.Hrs",
> "Mon.Hrs"), class = "data.frame", row.names =
integer(0))
That's not a valid dataframe, it's just labelled as one.  If you tried 
to create it with data.frame(), you'd get something different.

 > test <- data.frame(Code = NULL, Name = NULL, Address = NULL, Sun.Hrs 
= NULL,
+     Mon.Hrs = NULL)
 > test
data frame with 0 columns and 0 rows

You can create a zero-row dataframe as long as you put 0-length vectors 
in as columns.  NULL is not a vector.

 > test <- data.frame(Code = numeric(0), Name = numeric(0), Address = 
numeric(0), Sun.Hrs = numeric(0),
+     Mon.Hrs = numeric(0))
 > test
[1] Code    Name    Address Sun.Hrs Mon.Hrs
<0 rows> (or 0-length row.names)

If you do that, head() works:

 > head(test)
[1] Code    Name    Address Sun.Hrs Mon.Hrs
<0 rows> (or 0-length row.names)

So this is a bug in openxlsx.  It's also a well-known limitation of the 
S3 object system:  you can easily create things that are labelled with a 
certain class, but aren't valid objects of that class.

Duncan Murdoch