RE: [NMusers] No Fit file -> bug or not a bug?

Mark,

There is no need for a "platform-independent-Fortran based fix". (Because
it only happens on Windows). And as you point out, there already is
platform specific code that is distributed, nmfe6.bat, which could readily
be modified to fix the issue (at least for the control file). NMTRAN has to
read the control stream and the data file in order to create FCON and FDATA
anyway. It should not be a big deal to automatically add a line to the end
of each of these files.

The point is that it should not be something that requires a work-around or
that is just something the experienced know about. It should be fixed and a
non-issue. No?

Bill


-----Original Message-----
From: owner-nmusers_at_globomaxnm.com [mailto:owner-nmusers_at_globomaxnm.com] On
Behalf Of Mark Sale - Next Level Solutions
Sent: Wednesday, May 23, 2007 11:43 AM
Cc: 'nmusers'
Subject: RE: [NMusers] No Fit file -> bug or not a bug?

Bill,
 I'm not sure there is a platform-independent-Fortran based fix (Alison
would know better). You could edit nmfe6.bat to append a crlf, or go
to another language for your front end application, one with more text
processing flexibility. I assume it is part of everyone's QC to
confirm that the number of observations and individuals in the output
file matches what they expected.
BTW, I think nmfe6.bat and be "fixed" by:
Create a file c:\nmvi\util\end.txt
put in it:

; end

(note crlf before and after ; end - all you really need is a single
crlf)

add this text after the first line of c:\nmvi\util\nmfe6.bat

type c:\nmvi\util\end.txt >> %1


Mark Sale MD
Next Level Solutions, LLC
www.NextLevelSolns.com


> -------- Original Message --------
> Subject: RE: [NMusers] No Fit file -> bug or not a bug?
> From: "Bill Bachman" <bachmanw_at_comcast.net>
> Date: Wed, May 23, 2007 11:25 am
> To: "'Mark Sale - Next Level Solutions'" <mark_at_nextlevelsolns.com>
> Cc: "'nmusers'" <nmusers_at_globomaxnm.com>
>
> I guess this depends on how you define "bug". If a feature of an
> application (NONMEM run with Fortran) causes an error (failure of a
> table to
> be created, as in this case, or a data record not to be included in
> the data
> set) on one of the intended operating systems (Windows), I would (in my
> opinion) call it a bug. I would suggest fixing this known issue so
> others
> don't have to puzzle over it (or worse, omit a data record unknowingly).
>
> -----Original Message-----
> From: owner-nmusers_at_globomaxnm.com
> [mailto:owner-nmusers_at_globomaxnm.com] On
> Behalf Of Mark Sale - Next Level Solutions
> Sent: Tuesday, May 22, 2007 7:04 PM
> Cc: nmusers
> Subject: RE: [NMusers] No Fit file
>
> Thanks for clarifying this Alison, I think I did know this is a fortran
> I/O issue, not really a NMTRAN issue. Clearly we should all go back to
> MS-DOS edit ;-) (only us really old people even remember MS-DOS).
>
>
>
> Mark Sale MD
> Next Level Solutions, LLC
> www.NextLevelSolns.com
>
>
> > -------- Original Message --------
> > Subject: RE: [NMusers] No Fit file
> > From: "Alison Boeckmann" <alisonboeckmann_at_fastmail.fm>
> > Date: Tue, May 22, 2007 5:41 pm
> > To: "Mark Sale - Next Level Solutions" <mark_at_nextlevelsolns.com>
> > Cc: "nmusers" <nmusers_at_globomaxnm.com>
> >
> > This is not a bug in NM-TRAN. But I can see how it happens in
> > MS-DOS. Both wordpad and notepad (unlike ms-dos edit and unix vi) will
> > write a file whose last line does not terminate with cr/lf. Such a line
> > is invisible to the Fortran I/O libraries that are part of g77
> > (in the MS-DOS environment) and FORTRAN 77 5.0 (in the Solaris
> > environment).
> >
> > I tested this in both MS-DOS and Solaris with a little fortran program
> > that branches to statement 10 when an end of file condition occurs.
> >
> >
> > program read
> > character*10 text
> > read (5,55,end=10) text
> > print *,text
> > 55 format (A10)
> > stop
> > 10 print *,'nothing read'
> > end
> >
> > I created a file named infile containing only "1234567890" with no
> cr/lf
> > using notepad. I copied it to Solaris using ftp. In both environments,
> > when the program is run with infile as input, it produces
> > the output "nothing read". E.g., in Solaris,
> >
> > % f77 read.for
> > read.for:
> > MAIN read:
> > % a.out<infile
> > nothing read
> >
> > I don't know how to overcome this problem. Changing the format from A10
> > to (say)
> > A1 or * does not help. It seems to me it is a bug (or maybe a design
> > feature?) in the Fortran I/O library.
> >
> >
> > On Tue, 22 May 2007 12:56:18 -0700, "Mark Sale - Next Level Solutions"
> > <mark_at_nextlevelsolns.com> said:
> > > Navin, If you're not getting the table file you expect, if may be
> > > because NMtran (5 at least) seems to have a small bug that it will
> > > sometimes ignore the last line of a control file if there isn't a
> > > carriage return/line feed at the end. Try just hitting Enter after
> > > the $TABLE record and saving. WRT plots, two options are Xpose from
> > > the Upsala group (R/Splus, http://xpose.sourceforge.net) or an excel
> > > macro from Next Level Solutions (www.nextlevelsolns.com)
> > >
> > >
> > >
> > > Mark Sale MD Next Level Solutions, LLC www.NextLevelSolns.com
> > >
> > >
> > > > -------- Original Message -------- Subject: RE: [NMusers] No Fit
> > > > file From: "Bachman, William (MYD)" <bachmanw_at_iconus.com> Date:
> Tue,
> > > > May 22, 2007 2:48 pm To: "navin goyal" <navin1180_at_gmail.com>,
> > > > "nmusers" <nmusers_at_globomaxnm.com>
> > > >
> > > > according to your code, your run should have created a file named
> > > > "drug3.sdtab". this file contains a table with these columns "ID
> > > > TIME AMT MDV DV EVID CMT" and DV PRED RES WRES. You can import
> > > > this file into Excel. You will have to manually create the graphs
> > > > unless you have a macro that can do it.
> > > >
> > > > ________________________________
> > > >
> > > > From: owner-nmusers_at_globomaxnm.com [mailto:owner-
> > > > nmusers_at_globomaxnm.com] On Behalf Of navin goyal Sent: Tuesday, May
> > > > 22, 2007 2:20 PM To: nmusers Subject: [NMusers] No Fit file
> > > >
> > > >
> > > > Hello everybody, I am trying to run a control stream for my data
> > > > using NONMEM 6. I get the output file with the objective function
> > > > and other estimates. The problem is that I do not get the fit
> > > > file that can be opened in excel. I guess I am missing out
> > > > something in the control stream, but just cannot figure it out.
> > > > Any suggestions would be appreciated The last part in the control
> > > > stream is posted below
> > > >
> > > >
> > > > $ESTIMATION METHOD=1 MAXEVAL=9999 PRINT=5 POSTHOC $COVARIANCE
> > > >
> > > > $TABLE ID TIME AMT MDV DV EVID CMT NOPRINT ONEHEADER
> > > > FILE=drug3.sdtab
> > > >
> > > > thanks a lot
> > > >
> > > > Navin
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > > ___________________________________________________________________-
> > > > ____________________________________________________________________
> > > >
> > > >
> > > >
> > > >
> > > > This e-mail transmission may contain confidential or legally
> > > > privileged information that is intended only for the individual or
> > > > entity named in the e-mail address. If you are not the intended
> > > > recipient, you are hereby notified that any disclosure, copying,
> > > > distribution, or reliance upon the contents of this e-mail is
> > > > strictly prohibited. If you have received this e-mail transmission
> > > > in error, please reply to the sender, so that ICON can arrange for
> > > > proper delivery, and then please delete the message. Thank You.
> > >
> > --
> > Alison Boeckmann
> > alisonboeckmann_at_fastmail.fm

Received on Wed May 23 2007 - 12:50:40 EDT