On Jun 28, 2015, at 9:34 AM, Riccardo (Jack) Lucchetti <r.lucchetti(a)univpm.it&gt; wrote: Folks, it occurred to me that our way of communicating with R (and potentially other packages, but let's focus on R for the moment) could be much improved if we had R functions for reading/writing bundles as R lists. I started experimenting a little with the "XML" R package. For example: <hansl> bundle a = null a.n = 512 bwrite(a, "a.xml") <hansl> and then do <R> library(XML); a <- xmlParse("a.xml"); b <- xmlToList(a); c <- b[1]$`bundled-item`$text n <- as.numeric(c); print(n); </R> and this is quite cool. Of course this would have to be properly generalised/extended on the R side. I suppose that for someone who's fluent in R it shouldn't be too much work. Is anybody willing to help with this? ------------------------------------------------------- Riccardo (Jack) Lucchetti Dipartimento di Scienze Economiche e Sociali (DiSES) Università Politecnica delle Marche (formerly known as Università di Ancona) r.lucchetti(a)univpm.it http://www2.econ.univpm.it/servizi/hpp/lucchetti ------------------------------------------------------- _______________________________________________ Gretl-devel mailing list Gretl-devel(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-devel

On Jun 28, 2015, at 9:34 AM, Riccardo (Jack) Lucchetti <r.lucchetti(a)univpm.it&gt; wrote: Folks, it occurred to me that our way of communicating with R (and potentially other packages, but let's focus on R for the moment) could be much improved if we had R functions for reading/writing bundles as R lists. I started experimenting a little with the "XML" R package. For example: <hansl> bundle a = null a.n = 512 bwrite(a, "a.xml") <hansl> and then do <R> library(XML); a <- xmlParse("a.xml"); b <- xmlToList(a); c <- b[1]$`bundled-item`$text n <- as.numeric(c); print(n); </R> and this is quite cool. Of course this would have to be properly generalised/extended on the R side. I suppose that for someone who's fluent in R it shouldn't be too much work. Is anybody willing to help with this? ------------------------------------------------------- Riccardo (Jack) Lucchetti Dipartimento di Scienze Economiche e Sociali (DiSES) Università Politecnica delle Marche (formerly known as Università di Ancona) r.lucchetti(a)univpm.it http://www2.econ.univpm.it/servizi/hpp/lucchetti ------------------------------------------------------- _______________________________________________ Gretl-devel mailing list Gretl-devel(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-devel

From the bwrite doc it would seem that the dotdir is only used >when you add a third parameter, bwrite(a, "a.xml", 1). So my guess >is that the file goes into your current workdir instead (which is >user-configurable).

> Ah, right. Adding the 3rd argument successfully writes a.xml to dotdir, but it's still empty. >

Confirmed (also on Windows). Looks like a bug.

-sven

> From the bwrite doc it would seem that the dotdir is only used > >when you add a third parameter, bwrite(a, "a.xml", 1). So my > guess >is that the file goes into your current workdir instead (which > is >user-configurable). > Ah, right. Adding the 3rd argument successfully writes a.xml to > dotdir, but it's still empty.

Are you using totally current CVS/snapshot? There was a bug that >Jack spotted recently which could result in an empty file when >using bwrite(); it's now fixed.

Allin