During my troubleshooting of a failed NONMEM run, I want to confirm the
correct exclusion of some records (e.g., IGNORE CMT=2) for the input file.
However, I could not find the FDATA file in the run directory. Is there any
way to force NONMEM to generate a FDATA? I am also wondering about the
possible causes of missing FDATA.
Even the FDATA file is available, is it possible to include the column
heads (e.g., ID, DV) on the top of the numbers? So, I can confirm which
column is excluded or retained.
In addition, I am not sure whether the dataset related warnings (e.g., DATA
WARING 1, RECORD 100, DATA ITEM 11, THE DV IS NULL BUT MDV IS 0) is
refereeing to the FDATA file or the raw input CSV file. In other words, the
problematic record 100 represents the row 100 in the FDATA file or the CSV
file?
Thank you,
Mark
FDATA file missing
Hi Mark,
I'm guessing that you are probably running NONMEM via PsN. The default behavior
of PsN is to suppress some output files, including the FDATA file. To request
that the FDATA file be retained, you can include the option -extra_output=fdata
on your PsN execute command line.
I don't think there is any way to include the column headings (since they are
ignored and not actually read into NONMEM), but having the FDATA file should
help you troubleshoot your problem. As far as I know, the warnings relate to
the location in the FDATA file, again, because the ignored rows are not
actually being read in.
Best of luck,
Sarah
Sarah F. Cook, PhD
Adjunct Assistant Professor
University at Buffalo
Department of Pharmaceutical Sciences
E-mail: [email protected]
Phone: (570) 362-2981
Quoted reply history
________________________________
From: [email protected] <[email protected]> on behalf of
Mark Tepeck <[email protected]>
Sent: Tuesday, November 26, 2019 11:13 AM
To: [email protected]
Subject: [NMusers] FDATA file missing
During my troubleshooting of a failed NONMEM run, I want to confirm the correct
exclusion of some records (e.g., IGNORE CMT=2) for the input file. However, I
could not find the FDATA file in the run directory. Is there any way to force
NONMEM to generate a FDATA? I am also wondering about the possible causes of
missing FDATA.
Even the FDATA file is available, is it possible to include the column heads
(e.g., ID, DV) on the top of the numbers? So, I can confirm which column is
excluded or retained.
In addition, I am not sure whether the dataset related warnings (e.g., DATA
WARING 1, RECORD 100, DATA ITEM 11, THE DV IS NULL BUT MDV IS 0) is refereeing
to the FDATA file or the raw input CSV file. In other words, the problematic
record 100 represents the row 100 in the FDATA file or the CSV file?
Thank you,
Mark
Hello Mark,
in addition to Sarah's suggestions, you could run NONMEM via psn using
option clean=0 which will stop psn from cleaning the temp files generated
by NONMEM.
for the data issues, it refers to the actual data that's read into NONMEM
i.e after excluding all ignore statements and dropped columns, a workaround
is to filter out the dataset in the same way it's being filtered in the
NMtran control stream or kill the run before it crashes and you should have
the FDATA file in the NM folder.
Mahmoud
Quoted reply history
On Tue, Nov 26, 2019 at 7:10 PM Cook, Sarah <[email protected]> wrote:
> Hi Mark,
>
>
> I'm guessing that you are probably running NONMEM via PsN. The default
> behavior of PsN is to suppress some output files, including the FDATA file.
> To request that the FDATA file be retained, you can include the option
> -extra_output=fdata on your PsN execute command line.
>
>
> I don't think there is any way to include the column headings (since they
> are ignored and not actually read into NONMEM), but having the FDATA file
> should help you troubleshoot your problem. As far as I know, the warnings
> relate to the location in the FDATA file, again, because the ignored rows
> are not actually being read in.
>
>
> Best of luck,
>
> Sarah
>
>
>
> Sarah F. Cook, PhD
> Adjunct Assistant Professor
> University at Buffalo
> Department of Pharmaceutical Sciences
> E-mail: [email protected]
> Phone: (570) 362-2981
> ------------------------------
> *From:* [email protected] <[email protected]> on
> behalf of Mark Tepeck <[email protected]>
> *Sent:* Tuesday, November 26, 2019 11:13 AM
> *To:* [email protected]
> *Subject:* [NMusers] FDATA file missing
>
>
> During my troubleshooting of a failed NONMEM run, I want to confirm the
> correct exclusion of some records (e.g., IGNORE CMT=2) for the input file.
> However, I could not find the FDATA file in the run directory. Is there any
> way to force NONMEM to generate a FDATA? I am also wondering about the
> possible causes of missing FDATA.
>
> Even the FDATA file is available, is it possible to include the column
> heads (e.g., ID, DV) on the top of the numbers? So, I can confirm which
> column is excluded or retained.
>
> In addition, I am not sure whether the dataset related warnings (e.g.,
> DATA WARING 1, RECORD 100, DATA ITEM 11, THE DV IS NULL BUT MDV IS 0) is
> refereeing to the FDATA file or the raw input CSV file. In other words, the
> problematic record 100 represents the row 100 in the FDATA file or the CSV
> file?
>
> Thank you,
>
> Mark
>
--
*Mahmoud Tareq AbdelwahabPhD StudentPharmacometrics Research GroupDivision
of Clinical PharmacologyDepartment of MedicineUniversity of Cape TownK45
Old Main BuildingGroote Schuur HospitalObservatory, Cape Town 7925, South
Africa.*
*Tel: +27 21 650 4861Email: [email protected] <[email protected]>*
Hi Mahmoud and Sarah,
Thank you so much for your input. Both solutions do the job when I use Psn.
-extra_output=fdata adds fdata to the root folder, while clean=0 keeps
fdata in the Psn folder.
I hope that Robert J. Bauer could consider the column heading feature for
future NONMEM versions. I believe it would enable more efficient
troubleshooting whenever needed.
Best,
Mark
Quoted reply history
On Tue, Nov 26, 2019 at 9:22 AM Mahmoud Tareq Abdelwahab <
[email protected]> wrote:
> Hello Mark,
>
> in addition to Sarah's suggestions, you could run NONMEM via psn using
> option clean=0 which will stop psn from cleaning the temp files generated
> by NONMEM.
>
> for the data issues, it refers to the actual data that's read into NONMEM
> i.e after excluding all ignore statements and dropped columns, a workaround
> is to filter out the dataset in the same way it's being filtered in the
> NMtran control stream or kill the run before it crashes and you should have
> the FDATA file in the NM folder.
>
> Mahmoud
>
> On Tue, Nov 26, 2019 at 7:10 PM Cook, Sarah <[email protected]> wrote:
>
>> Hi Mark,
>>
>>
>> I'm guessing that you are probably running NONMEM via PsN. The default
>> behavior of PsN is to suppress some output files, including the FDATA file.
>> To request that the FDATA file be retained, you can include the option
>> -extra_output=fdata on your PsN execute command line.
>>
>>
>> I don't think there is any way to include the column headings (since they
>> are ignored and not actually read into NONMEM), but having the FDATA file
>> should help you troubleshoot your problem. As far as I know, the warnings
>> relate to the location in the FDATA file, again, because the ignored rows
>> are not actually being read in.
>>
>>
>> Best of luck,
>>
>> Sarah
>>
>>
>>
>> Sarah F. Cook, PhD
>> Adjunct Assistant Professor
>> University at Buffalo
>> Department of Pharmaceutical Sciences
>> E-mail: [email protected]
>> Phone: (570) 362-2981
>> ------------------------------
>> *From:* [email protected] <[email protected]> on
>> behalf of Mark Tepeck <[email protected]>
>> *Sent:* Tuesday, November 26, 2019 11:13 AM
>> *To:* [email protected]
>> *Subject:* [NMusers] FDATA file missing
>>
>>
>> During my troubleshooting of a failed NONMEM run, I want to confirm the
>> correct exclusion of some records (e.g., IGNORE CMT=2) for the input file.
>> However, I could not find the FDATA file in the run directory. Is there any
>> way to force NONMEM to generate a FDATA? I am also wondering about the
>> possible causes of missing FDATA.
>>
>> Even the FDATA file is available, is it possible to include the column
>> heads (e.g., ID, DV) on the top of the numbers? So, I can confirm which
>> column is excluded or retained.
>>
>> In addition, I am not sure whether the dataset related warnings (e.g.,
>> DATA WARING 1, RECORD 100, DATA ITEM 11, THE DV IS NULL BUT MDV IS 0) is
>> refereeing to the FDATA file or the raw input CSV file. In other words, the
>> problematic record 100 represents the row 100 in the FDATA file or the CSV
>> file?
>>
>> Thank you,
>>
>> Mark
>>
>
>
> --
>
>
>
>
>
>
>
>
>
>
> *Mahmoud Tareq AbdelwahabPhD StudentPharmacometrics Research GroupDivision
> of Clinical PharmacologyDepartment of MedicineUniversity of Cape TownK45
> Old Main BuildingGroote Schuur HospitalObservatory, Cape Town 7925, South
> Africa.*
>
>
> *Tel: +27 21 650 4861Email: [email protected] <[email protected]>*
>
>