Bug #1039
closedcreate and revise schema docs for ecogrid messages
0%
Description
Schema document for query, read, RSIDS, structure, with definitions and
examples. This is a formalization of the schemas we developed at the Seattle
EcoGrid meeting, meant to be captured and documented.
Target deadline: April 18
Updated by Matt Jones over 21 years ago
Query document schema is largely complete and is in CVS (needs docuemntation and
the operator list enumeration).
Need to continue work on message structure for the result set (RSIDS).
Updated by Matt Jones over 21 years ago
Added a resultset schema document (resultset.xsd) to CVS. It is an initial pass
at a structure for a result set. Still needs work, as does the query.xsd, but
both are a good basis for discussion at the meeting next week in San Diego.
Updated by Matt Jones over 21 years ago
Modify both query.xsd and resultset.xsd to accomodate the changes we discussed
at the meeting. This may incorporate XQuery.
Who: Matt
Target: May 26
Updated by Jing Tao over 21 years ago
To my understanding, resultset.xsd is for RISD type in queryInterface.wsdl. If
it has query itself (filter) in resultset schema, it will be very easy for
PortType decodeRISD. Because RISD already has the query part.
By the way, resultset has an element name "sendTime" and in
queryInterface3.wsdl, RISD type has an element named "timeStamp". Are they
redundant?
Updated by Matt Jones about 21 years ago
At the Seattle (Sept 2003) EcoGrid meeting we decided:
Revise query.xsd:
-- add repeatable returnfield element for specifying structure of return records
Revise resultset.xsd:
-- add system, namespace, modificationDate, and creationDate to each record
-- identifiers for a record are considered relative URIs, with a base value
of the "system" attribute for that metadata, where the system should
contain the GSH for an ecogrid enabled server:
so, for example, system="ecogrid://ecogrid.nceas.ucsb.edu"
identifier="/knb.1.1"
-- modify the record internal structure by creating a composite document
that contains all of the values specified in returnfield elements with
their parent elements. If the path does not reference a leaf node
text element, then all of the children of that node are returned.
For example, if the query had:
returnfield="/eml/dataset/title" and returnfield="/eml/dataset/creator"
The record would look like:
<record ...>
<eml>
<dataset>
<title>Some random title</title>
<creator>
<individualName>
<givenName>Matthew</givenName>
<surName>Jones</surName>
</individualName>
</creator>
</dataset>
</eml>
Updated by Matt Jones almost 19 years ago
Initial implementation based on globus finished. Closing these bugs as we move
to a new implementation based on axis.