[R] read.xlsx function crashing R Studio

[Rainer Hurling]

Hi Kevin,

Am 21.08.2016 um 19:30 schrieb Kevin Kowitski:
> Hey everyone, 
> 
>    I have used read.xlsx in the past rather than XLConnect for importing
> Excel data to R.  However, I have been finding now that the read.xlsx
> function has been causing my R studio to Time out.  I thought it might
> be because the R studio I had was out of date so I installed R studio
> X64 3.3.1 and reinstalled the xlsx package but it is still failing.  I
> have been trying to use XLConnect in it's place which has been working,
> excpet that I am running into memory error:
>               Error: OutOfMemoryError (Java): GC overhead limit exceeded
>       
> I did some online searching and found an option to increase memory:
>               "options(java.parameters = "-Xmx4g" )
> 
> but it resulted in this new memory Error:
> 
>              Error: OutOfMemoryError (Java): Java heap space
> 
> Can anyone provide me with some help on getting the read.xlsx function
> working?
> 
> -Kevin

There are interesting alternatives with other packages, as mentioned
before by Jim Holtman and Hadley Wickham.

If there are serious reasons to use the xlsx package, I had success with
the following, somewhat ugly workaround:


# Increase before package loading
options(java.parameters = "-Xmx8000m")
# Java garbage collection function
jgc <- function() .jcall("java/lang/System", method = "gc")
library(xlsx)
# if you like to use ISO dates
options(xlsx.date.format="yyyy-mm-dd")
# ATTENTION:  Loading of package xlsx changes decimal point
# to comma in a German locale
Sys.setlocale(category = "LC_NUMERIC", locale="C")  # UGLY HACK !!!


# Later in your code, i.e. before using addDataFrame(), use
jgc()


HTH,
Rainer Hurling