Re: detrend keywords

From: JJ Kavelaars <jjk_at_hia-iha.nrc-cnrc.gc.ca>
Date: Wed, 11 Apr 2007 14:42:31 -0700

Hi,

I looked through the failed transfer logs for TP-> CADC transfers.
Here are the 2 types of failures we have been finding:

A file that failed fitsverify at our end is 695973p_weight.fits

this is the weight map image for 695973p

the failure is caused by the use of 'e' instead of 'E' in the CD
matrix values for this image, specifically CD1_1 and CD2_2 which have
the values -5.194e-05 and 5.194e-05

I retrieved the image 695973p from our system and don't see the lower
case 'e' problem in that file so I don't know how that can get into the
weight header.

To fix this TP should modify the CD1_1 and CD2_2 values to be
compliant with the standard.



another example of a failed file is 803070p_weight where the keyword
DATEEND has an invalid value (Null).

in this case this IS an image that was transferred to CADC with the
string 'Null' in the value field for DATEEND. This was done via the
'no-check' path from CFHT to CADC and was implemented to clear a
transfer backlog from CFHT to CADC. We have a small, but non-zero,
subset of such files.

I'm not sure what to do here. Likely just deleting DATEEND from the
headers of the weights would be sufficient. The value is not
applicable to those images anyway.

Other suggestions?

JJ




On 11-Apr-07, at 1:33 PM, mellier_at_iap.fr wrote:

>
> Dear Kanoa
>
> Ok I will do this with FRed.
>
> Cheers
>
> yannick
>
>> Thanks for facilitating the meeting this morning JJ!
>>
>> I took a look at the detrend headers in our repository and I don't
>> see
>> the lower-case "e" problem we were talking about earlier.
>> Therefore, I
>> just wanted to remind JJ that you said you would provide an
>> example and
>> that Yannick said you would provide a list of affected files.
>> Actually
>> just one or two affected file names should be enough to trace the
>> problem.
>>
>> Thanks!
>>
>> -Kanoa
>>
>
>
>
Received on Wed Apr 11 2007 - 11:42:55 HST