Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
fread: showProgress must be 0 or 1 #1111
What I expected
What I did
install_github("Rdatatable/data.table") write.csv(1, "tmp.csv") fread("tmp.csv", verbose = TRUE)
# Error in fread("tmp.csv", verbose = TRUE) : # showProgress must be 0 or 1, currently
Works normally on Ubuntu 14.04 / R 3.1.0
Just moved to Win 7 64bit, after installing latest versions, am getting the same issue, but only when I run the 64bit version of R - the same code works in the 32bit version:
64-bit: showProgress Error
32-bit: works as expected
Can't reproduce either. Best guess is at some corruption in the 32bit/64bit DLLs on your Windows box. It has happened before and I've seen odd problems like this when an older version of a DLL gets loaded with newer R package code.
I looked at the code as well and can't see any problems. When data.table starts it creates a global option. Type the following. This is what I get and is correct.
> require(data.table) > getOption("datatable.showProgress")  1 > storage.mode(getOption("datatable.showProgress"))  "integer"
When you call
fread = function(..., showProgress=getOption("datatable.showProgress"), ...)
then all it does is pass that down to C level, and wraps with
Now because that argument is the last one, that's why I think some mismatch in DLL versions has happened i.e. it is somehow calling an old DLL that has less arguments. I know this sounds like Windows bashing but it's all I can think of as I've seen it before.
Regardless, I've added extra tracing at C level to hone in on the problem :
Closing for now but please let us know whether either the reboot-purge-reinstall or the new tracing reveals anything.
Sorry for the delayed follow up. I have another windows system that had the same issue - Win 8.1 64 bit, remainder of the setup was the same. System reboots didn't help, but uninstalling/reinstalling data.table (again, from github) twice on each machine resolve the issue. Unfortunately, I have no output to relay as everything seems to work properly. If you'd like me to repost any system info (or additional system info), please let me know. Thanks for data.table in general and the help on this one, Matt!
I had the same problem, only after updating to the development version 1.9.5. The error persists even if
@HughParsonage did you tried the way which helps ericsgagnon to solve the issue? I would recommend to close all R session before reinstalling on clean session. I think it may happens some session were not closed correctly so the process manager may be useful to confirm that. Then reinstall on clean session should be enough. I wonder if using
I second the same error when using 1.9.7. I down graded to 1.9.6 and it works fine. My best guess is that the file has a number of description lines akin to a report.
where "PROVIDER NAME"... is the start of the actual table.
referenced this issue
Nov 30, 2016
added a commit
Nov 30, 2016
attached base packages:
other attached packages:
loaded via a namespace (and not attached):