I would like to be able to determine the number of frames in an image during a query. Upon closer examination of the Dicom specs however, it seems that, at most, only tags specified as Unique or Required keys are to be returned from a query, and optionally the Optional keys.
Hence it appears that tags not specified as keys cannot be obtained without retrieving the image itself. Is this correct? I'm surprised by this sort of explicit limitation in the Dicom query model. Any idea of the reason for this limitation? The concept of "keys" in most database models relates to the manner in which data "records" are uniquely identified, but implies no restriction on availability of non key values from querries.
Restrictions in the Dicom query model
Yes, your assumption is correct.
The Number of Frames tag (0028,0008) is not one of the required keys at the Image query level, so it'll be up to the Query SCP application to return any useful information, assuming the Query SCU application actually sends this tag in the C-FIND request, because this key is not required for either the C-FIND SCU or the SCP applications. Whether the SCU/SCP supports this optional tag or not can be found from their Dicom conformance statements.
The Number of Frames tag (0028,0008) is not one of the required keys at the Image query level, so it'll be up to the Query SCP application to return any useful information, assuming the Query SCU application actually sends this tag in the C-FIND request, because this key is not required for either the C-FIND SCU or the SCP applications. Whether the SCU/SCP supports this optional tag or not can be found from their Dicom conformance statements.
Only if that tag is listed in the Supported Attribute Key list for C-FIND command in the Dicom conformance statements for both the query SCU and SCP applications, i.e., the C-FIND SCU needs to send that tag in the C-FIND request, and the C-FIND SCP knows how to return useful information for that tag (instead of just ignoring it)