[JCR-3487] Provide a utility to run JCR queries adopting a fluent-interfaces approach

• Type: Improvement
• Status: Open
• Priority: Minor
• Resolution: Unresolved
• Component/s: jackrabbit-jcr-commons
• Assignee: Unassigned
• Reporter: Simone Tripodi
• Created: 17/Dec/12 20:06
• Updated: 17/Dec/12 20:06

Description

(moving the discussion from OAK-529)

I propose an approach that, without altering the current codebase, but rather adding a new package on top of existing APIs, tries to simplify the query building and execution, i.e. rather than coding the following:

QueryManager qm = session.getWorkspace().getQueryManager();
Query q = qm.createQuery("select text from [nt:base] where id = $id", Query.JCR_SQL2);
q.bindValue("id", vf.createValue("1"));
QueryResult r = q.execute();

String text = null;

RowIterator it = r.getRows();
if (it.hasNext()) {
    Row row = it.nextRow();
    text = row.getValue("text").getString();
}

users could write the same sentence using a shortcut like

String text = on(qm)
              .sql2Query("select text from [nt:base] where id = $id")
              .where("id").is("1")
              .execute(textHandler);

where textHandler is:

QueryResultHandler<String> textHandler = new QueryResultHandler<String> {

    @Override
    public String handle(QueryResult queryResult)
            throws RepositoryException {

        RowIterator it = queryResult.getRows();
        if (it.hasNext()) {
            Row row = it.nextRow();
            return row.getValue("text").getString();
        }

        return null;
    }

}

the QueryResultHandler should simplify reducing redundant query handling code, improving reusability.