[kepler-users] passing multiple data frames between RExpression actors

[Scott Porter]

I am in 1.0 on OS X.  I don't feel up to installing the dev version from the SVN right now.
Maybe later when I have more time.

I did figure out a workaround in R that doesn't involve saving the workspace... 
I can pass a list of dataframes. 

It's not ideal, since I have to know the name of the list elements I care about
and check for them manually in the next Rexpression actor, but I think it 
will work for me as a stopgap as it is pretty flexible and seems to work reliably.

Thanks for confirming the bug for me.

Scott

-----Original Message-----
From: Derik Barseghian [mailto:barseghian at nceas.ucsb.edu]
Sent: Fri 11/13/2009 2:01 PM
To: Scott Porter; kepler-users at kepler-project.org
Subject: Re: [kepler-users] passing multiple data frames between RExpression actors
 
Hi Scott,

Are you using Kepler 1.0? I tested this with a simple workflow here,  
and it seems to be a bug in 1.0, but not in the current development  
version of Kepler in SVN. (I got two copies of the same dataframe in  
my second Rexpression actor).
Saving the workspace between Rexpression actors is one solution, but -- 
save is a feature we try very hard to avoid using since it subverts  
Kepler -- tokens aren't flowing along relations when it's used, and  
can lead to confusion in later Rexpression actors where it's unclear  
what variables are available. However you might use this as a stopgap  
until we release 2.0, or if you're comfortable with checking out and  
building Kepler from SVN, and willing to deal with things frequently  
breaking and changing as we work towards 2.0, you could try the  
development version of Kepler, the instructions for which are here:
https://kepler-project.org/developers/teams/build/documentation/build-system-instructions

Derik

On Nov 12, 2009, at 11:19 PM, Scott Porter wrote:

>
> I have a workflow where everything works.  I can pass a single data  
> frame with no problem between two RExpression actors.  However, if I  
> add a port for a second dataframe, the first dataframe port now  
> passes junk data from some random memory location (in some cases a  
> different dataframe, in other cases a random function, etc).  Is it  
> a known limitation of the RExpression actor that it can only pass  
> one dataframe?  Is the only workaround to pass an R workspace  
> between the actors?
>
> _______________________________________________
> Kepler-users mailing list
> Kepler-users at kepler-project.org
> http://mercury.nceas.ucsb.edu/kepler/mailman/listinfo/kepler-users



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.nceas.ucsb.edu/kepler/pipermail/kepler-users/attachments/20091113/0c447cd8/attachment.html>