[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: [OASIS Issue Tracker] (OSLCCORE-111) Query spec and relative operators for string literals
[ https://issues.oasis-open.org/browse/OSLCCORE-111?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] David Honey updated OSLCCORE-111: --------------------------------- Proposal: The spec should specify that support for ">", ">=", "<", and "<=" on string properties: 1) is a MAY (and not the currently implied MUST). 2) Where supported, the semantics are server dependent. Applications using RDBs might perform string ordering using locale-dependent collators. The spec should be persistence-agnostic. 3) where not supported, specify that a server SHOULD return a 501 "Not Implemented" response. Because the spec was silent on this before, existing implementations might have used other error responses. , and where not supported, and allow an implementation to fail if such a query is attempted. If an implementation supports those operators, I think the spec should be persistence-layer agnostic about the ordering of strings. was: The spec should exempt support for ">", ">=", "<", and "<=" on string properties and allow an implementation to fail if such a query is attempted. If an implementation supports those operators, I think the spec should be persistence-layer agnostic about the ordering of strings. > Query spec and relative operators for string literals > ----------------------------------------------------- > > Key: OSLCCORE-111 > URL: https://issues.oasis-open.org/browse/OSLCCORE-111 > Project: OASIS OSLC Lifecycle Integration Core (OSLC Core) TC > Issue Type: Task > Components: Query > Reporter: David Honey > Assignee: James Amsden > > https://open-services.net/bin/view/Main/OSLCCoreSpecQuery says for comparison_op: > "Semantics of datatypes and operations on these datatypes MUST work as defined in the SPARQL Query Language (reference: SPARQL). " > The problem is that for string properties, querying for ">", ">=", "<", and "<=" may be unimplementable, or the ordering might be determined by an RDB vendor and different from the SPARQL spec. The spec places too high a burden, and assumes an RDF persistence layer, on implementations. For example, some RDB vendors might use Locale-specific collators. -- This message was sent by Atlassian JIRA (v6.2.2#6258)
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]