Re: [NMusers] PSN vsWings

From: William Denney <wdenney_at_humanpredictions.com>
Date: Thu, 27 Jun 2019 07:12:29 -0400

Hi Andre,

I’m not totally clear what the issue is, but I can help eliminate so=
me of the questions:

The original source of the data before generating the NONMEM-ready dataset i=
s not important. The formatting of the NONMEM-ready dataset is vitally impo=
rtant.

My guess is that your dataset includes extraneous quotation marks as created=
 by the R command write.csv() (note the dot in the function name). I would g=
uess that if you use write_csv() from the readr library (note the underscore=
 in the function name), it will work because write_csv() does not add extran=
eous quotation marks.

For not being able to find the data, I don’t know how Wings differs f=
rom PsN there, but I’d check that the data file path is correct rela=
tive to the NONMEM control stream. And, I would run PsN from the same direc=
tory as the control stream.

Thanks,

Bill

> On Jun 27, 2019, at 06:12, Andre Jackson <jacksonan1945_at_gmail.com> wrote:
>
> I have a data set for which I have experienced the following:
>
> A Nonmem data set was prepared from SAS/R from an *.xpt file; however, whe=
n I tried to run under PSN it returned that it couldn’t find the dat=
a set.
> Using a data set that I had previously prepared with SAS and run under PSN=
 (i.e., not from an *.xpt file), I manually added additional subjects with t=
he same sampling scheme and the data set ran under PSN. However when I look=
ed at the outputted fit file only the original subjects had been read in and=
 parameters estimated for them by PSN;
> When I run estimation with the same data set under Wings it runs okay.
>
> Has anyone had a similar experience with data from an *.xpt file and if so=
 how can it be addressed?
>
>

Received on Thu Jun 27 2019 - 07:12:29 EDT

This archive was generated by hypermail 2.3.0 : Fri Sep 27 2019 - 17:05:28 EDT