datgnom read error

Interactive and automated data processing tools (PRIMUS, GNOM, AUTORG).
Scattering from simple bodies (BODIES), peak analysis (PEAK), data plotting (SASPLOT) etc.
Post Reply
Message
Author
rrock
Member
Posts: 4
Joined: 2014.10.24 18:04

datgnom read error

#1 Post by rrock » 2014.10.24 18:14

Hello,
I am running into a problem reading my averaged and background-subtracted datasets into datgnom (either from the command line, or through the primus interface). The error message is:

LOADATF --E- Error I/O with <filename>

I do not see anything obviously wrong with the data file, and the plotted curves look correct to me. I trimmed a few lines that had zero or negative values (due to noise) at either end of the data range, but that did not help. Nor did changing the line endings. I do not know how to proceed.

The datgnom versions are

command line:
datgnom (r3707)
Copyright (c) ATSAS Team, EMBL, Hamburg Outstation, 2010-2012

bundled with primus:
datgnom (r3791)
Copyright (c) ATSAS Team, EMBL, Hamburg Outstation, 2010-2012

This is on a Mac, but I have run into the same problem from a windows machine/primus.

I have also attached a data file that has the problem

Thanks
Attachments
full-bs.dat
(33.15 KiB) Downloaded 255 times

User avatar
AL
Administrator
Posts: 686
Joined: 2007.08.03 18:55
Location: EMBL Hamburg, Germany
Contact:

datgnom doesn't like zero errors

#2 Post by AL » 2014.10.27 12:36

Your data has zero errors (third column). If you don't have experimental errors you may remove the third column and datgnom will work.

rrock
Member
Posts: 4
Joined: 2014.10.24 18:04

Re: datgnom read error

#3 Post by rrock » 2014.10.27 23:22

Thanks-- that solved my issue.

I also now know why we were missing the experimental errors: the raw data files had six columns, but we only need the first three. A simple `cut -f1-3` gave us files that had appropriate errors after DATAVER / DATOP.

User avatar
AL
Administrator
Posts: 686
Joined: 2007.08.03 18:55
Location: EMBL Hamburg, Germany
Contact:

Re: datgnom read error

#4 Post by AL » 2014.11.05 16:48

rrock wrote:...the raw data files had six columns...
I'm glad it helped.
Where do these data come from?

rrock
Member
Posts: 4
Joined: 2014.10.24 18:04

Re: datgnom read error

#5 Post by rrock » 2014.11.20 22:52

Sorry that I missed your question-- they come from APS, at beamline 12-ID-B. No one around here has any idea what is found in the extra three columns of data, and there are no headers.

Cornelius
Member
Posts: 3
Joined: 2017.07.22 18:49

Re: datgnom doesn't like zero errors

#6 Post by Cornelius » 2017.07.22 19:07

AL wrote:Your data has zero errors (third column). If you don't have experimental errors you may remove the third column and datgnom will work.
Hi,

I also have raw data (after buffer substraction and averaging) without errors (third column). Is there a way to calculated it in order to plot a curve with I/sigma(I)=f(s) and be able to deposit the data in Biosis or Sasbdb?

Thanks in advance

User avatar
AL
Administrator
Posts: 686
Joined: 2007.08.03 18:55
Location: EMBL Hamburg, Germany
Contact:

SASBDB accepts data without errors

#7 Post by AL » 2017.07.23 22:31

Cornelius wrote:I also have raw data (after buffer substraction and averaging) without errors (third column). Is there a way to calculated it in order to plot a curve with I/sigma(I)=f(s) and be able to deposit the data in Biosis or Sasbdb?
You may deposit your data to SASBDB even without the experimental errors column.

Cornelius
Member
Posts: 3
Joined: 2017.07.22 18:49

Re: SASBDB accepts data without errors

#8 Post by Cornelius » 2017.07.24 08:18

AL wrote:You may deposit your data to SASBDB even without the experimental errors column.
Thanks, how about the plot I(s)/sigma(I)=f(s)?
We have been asking to provide it for publication, how is it relevant?

User avatar
AL
Administrator
Posts: 686
Joined: 2007.08.03 18:55
Location: EMBL Hamburg, Germany
Contact:

Re: how about the plot I(s)/sigma(I)=f(s)?

#9 Post by AL » 2017.07.27 10:01

Cornelius wrote:Thanks, how about the plot I(s)/sigma(I)=f(s)?
We have been asking to provide it for publication, how is it relevant?
Sorry, no idea.

Cornelius
Member
Posts: 3
Joined: 2017.07.22 18:49

Re: how about the plot I(s)/sigma(I)=f(s)?

#10 Post by Cornelius » 2017.07.27 14:01

Sorry, no idea.
Ok, Thanks anyway.

Post Reply