The CQL sorting proposal at
http://zing.z3950.org/cql/sorting.html
which was accepted as part of SRU/CQL 1.2 at the Hague meeting,
proposes the following XCQL representation:
<searchClause>
<index>dc.title</index>
<relation><value>=</value></relation>
<term>fish</term>
<sortKeys>
<key>
<index>dc.creator</index>
<modifiers>
<modifier>
<name>sort.missingValue</name>
<value>frog</value>
</modifier>
<modifier>
<name>sort.descending</name>
</modifier>
</modifiers>
</key>
</sortKeys>
</searchClause>
So far as I can tell, that's the only XCQL that's ever been proposed
for CQL's "sortby" -- I couldn't find anything on the v1.2 site at:
http://www.loc.gov:8081/standards/sru/
However, comparing this with the XCQL schema at:
http://www.loc.gov/standards/sru/xml-files/xcql.xsd
I see that the proposed sortby representation's handling of index
modifiers is inconsistent with the existing schema's handling of
relation modifiers and boolean modifiers, in that it uses <name>
rather than <type>, and omits <comparison>.
So I would like to propose a change to the "sortby" XCQL to bring its
index modifiers in line with the XCQL representation of CQL's other
modifiers, like so:
<searchClause>
<index>dc.title</index>
<relation><value>=</value></relation>
<term>fish</term>
<sortKeys>
<key>
<index>dc.creator</index>
<modifiers>
<modifier>
<type>sort.missingValue</type>
<comparison>=</comparison>
<value>frog</value>
</modifier>
<modifier>
<type>sort.descending</type>
</modifier>
</modifiers>
</key>
</sortKeys>
</searchClause>
Any objections? If not, I'll modify the proposal document, since
AFAIK that is currently still the only comprehensive public statement
of CQL sorting. (By the way, feel free to adopt any useful parts
document into the v1.2 specification.)
_/|_ ___________________________________________________________________
/o ) \/ Mike Taylor <[log in to unmask]> http://www.miketaylor.org.uk
)_v__/\ Among palaeontology grad students with dissertations to write,
side-projects are what makes the world go round.
|