Hi all,
so I was wondering what the general opinion is on handling Type 1C attributes in case their condition are not met.

Talking for example the PixelAspectRatio (0028,0034) considering an VL Endoscopic Image IOD: So when I know that my images/videos alsways have fixed PAR of 1:1, what would the right choice be:

1. Do not add the attribute to the SOP instance (or remove it)
2. Leave it within SOP instance with Zero Length value
3. Just set it with a value of "1\\1" anyhow

Considering "When the specified conditions are not met, Type 1C elements shall not be included in the Data Set." - from 7.4.2 Type 1C Conditional Data Elements, I would conclude that option (1.) is the correct one and that the attribute should not be included within the dataset at all. But how would you rate the other two options, would you say that in this case are generell wrong/bad practice or could they be fine too?

Kind regards,
Morty

You actually caught the appropriate passage in the standard and "shall not be included" is to be taken literally - you would violate the standard by including it.
Where it is left to the implementor's choice to include it although the condition is not met, the attribute description in the module table clearly states that by extending the condition with the phrase "may be present otherwise".

So ((1C) AND (NOT "may be present otherwise")) = don't include it.

I think that the second option is the worst because this is a very clear standard violation - 1C attributes must never have zero length values.
Option 3 would probably not cause any trouble, but this does change the fact that it is incorrect.