iraf.net

It appears that choosing "multispec" vs "echelle" for the format parameter in the apall task can produce different APNUM keywords in the output headers. I tested this on an image after unlearning apall (actually the imred.kpnoslit apall) and the relevant differences areAPNUM1 = '1 1 62.37 75.85'
APNUM2 = '2 2 152.79 166.42'vsAPNUM1 = '1 1 62.09 75.53'
APNUM2 = '2 2 153.01 166.66'According to the help file, the only difference between the two methods is the file extension. So the first question is why are the different and the second question is if there's a bug in one (or both) of the two calculations.I can send the exact file needed to reproduce the error upon request however I suspect it won't be hard to reproduce. Jason

Jason,I'll let Frank reply in detail next week when he returns, but I'm wondering whether there was some residual stuff in memory between runs. For example, if you run once with multispec then type 'flpr' and try again with echelle do the values still change?? If they are the same. there may be something that isn't initialized properly and without the 'flpr' you're seeing something leftover from the first run.Another test is to run the image twice but leave the format the same, both with and without and intervening 'flpr'. The results should be identical, if not then we have a (serious) problem.-Mike

[quote:3614b1fce8="fitz"]Jason,
Another test is to run the image twice but leave the format the same, both with and without and intervening 'flpr'. The results should be identical, if not then we have a (serious) problem.-Mike[/quote:3614b1fce8]As originally stated the problem was a false alarm. I didn't delete the database directory first between the runs. The APNUMs are only different if it is kept. However, it's not clear to me that keeping the database directory [i:3614b1fce8]should[/i:3614b1fce8] make a difference in the output since my references parameter is blank and the find parameter is "yes". So, it seems to me like for the default parameters of apall (i.e, after an unlearn) the output files should not be affected by previous contents of the database directory. I may just be being nearsighted though.Jason