ecl> ccdproc test389.fits fringeco+ illum=Illum30 fringe=Fringe
Warning: HDMGETR: No value found

• Adding the keyword FRINGSCL to the fringe frame did not help no matter whether the value was in integer or float.
  
• Changing values for EXPTIME from integer to float (e.g. 300.0) for all involved frames (testdata, Fringe, Illum) did not help
  </ul>
  I also took a closer look at the instrument file we use (we call it ccd.dat).
  The mapping for the EXPTIME keyword was like this:
  

  exptime exposure

  
  I then added the default value "1.0" as suggested in the thread quoted in the above posting. So the changed line looks like this:
  

   exptime exposure 1.0

  
  
  This solved the problem!
  
  My questions is why, though! None of my files had the exptime keyword missing, and I manually changed the values to be in float (i.e. 300.0, which gets written as 300.).
  Maybe the ccdproc, fringecor task creates a meta FITS file that lacks the exptime keyword, value for it, or has a non float format?
  
  Is this a possible bug or am I overlooking something?
  
  ----------- maybe move this part to bug section -------------------------------------
  
  On another note: Reading through a lot of help content, I noted that 'help mkfringecor' shows incorrect USAGE information and that the examples are incorrect as well.
  It seems mixed up with mkskycor:
  
  

  
  USAGE
      mkfringecor input output
  

  
  
  Is not working for me. The task expects "output=xy.fits"
  
  

  
  EXAMPLES
      1. The two examples below make an fringe  correction  image  from  a
      blank  sky  image, "sky017".  In the first example a separate fringe
      image is created and in the second the  fringe  image  replaces  the
      sky image.
      
          cl> mkskycor sky017 Fringe
          cl> mkskycor sky017 frg017
  

  
  
  It should say:
  'mkfringecor sky017 output=Fringe'
  'mkfringecor sky017 output=sky017'
  
  The second example is not self consistent as even with the different USAGE, the output filename should still be sky017 and not frg017, at least to my humble understanding of things.
  
  Thanks, and I will be checking back for comments and thoughts :-) </div> --- **Francisco Valdes** wrote on Aug 29, 2013
  Hi,
  
  The HDMGETR message comes from the keyword translation mechanism which means the translation file, ccd.dat in your case. I may have misunderstood but it sounded like you said your data has the keyword EXPTIME. The translation file you have says, instead, that it should look for the keyword EXPOSURE. You may be confused by the first part of the translation:
  

  exptime exposure

  
  The first word is the name that CCDPROC refers to the logical quantity of an exposure time. This is NOT the keyword it will look for unless there is no translation. With the translation it thinks, if I want the exposure time I will look for the keyword EXPOSURE.
  
  By putting in a default value the error went away because the exposure time it will use when it can't find EXPOSURE in the header is 1s. For fringe scaling it may be ok because it will end up using the ratio 1/1; i.e. no exposure time scaling.
  
  So the real solution, if I understood that you actually have EXPTIME and not EXPOSURE in your header, is to remove the exposure time translation from ccd.dat (or use "exptime exptime" which is redundant but should work).
  
  Frank Valdes</em>
  --- **antares** wrote on Aug 29, 2013
  Hi,
  
  Thank you very much for clarifying this. I will indeed get rid of the line, or make a 'dummy' for the same keyword in this case.
  
  The fringe correction worked not too badly, but I will try again with correct keyword mapping (or no mapping) and compare the results, to be sure.
  
  Thanks again,
  antares
  --- **antares** wrote on Aug 29, 2013
  I think I might have been not the only one in our group to misunderstand the mapping. So I am eager to read-up and maybe produce a template file that has some comments.
  
  Could you maybe point me to the correct direction as where to find documentation about the translation files.
  
  Is there a way to allow comments? # I.e. hash or other escaping?
  
  Thanks!
  --- *Last post on Aug 29, 2013*