[OPENJPA-2729] problem with ToUpperCase and ToLowerCase

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 2.4.2
• Component/s: query
• Environment: MySQL
• Assignee: Unassigned
• Reporter: Patrice DUROUX
• Created: 06/Mar/18 09:29
• Updated: 18/Oct/20 18:48

Description

Hi,

An exception is thrown in the following method (here the one of ToLowerCase):

when the getValue() is a PCPath and ((PCPath) getValue()).getFieldMapping(state).getColumns()[0])

throws:

java.lang.ArrayIndexOutOfBoundsException: 0{{ org.apache.openjpa.jdbc.kernel.exps.ToLowerCase.appendTo(ToLowerCase.java:47) org.apache.openjpa.jdbc.kernel.exps.MatchesExpression.appendTo(MatchesExpression.java:97) org.apache.openjpa.jdbc.kernel.exps.AndExpression.appendTo(AndExpression.java:59)}}

 

 

{{    public void appendTo(Select sel, ExpContext ctx, ExpState state, }}
        SQLBuffer buf, int index) {

        DBDictionary dict = ctx.store.getDBDictionary();
        String func = dict.toLowerCaseFunction;
        dict.assertSupport(func != null, "ToLowerCaseFunction");
        if (getValue() instanceof PCPath) {
            func = dict.getCastFunction(getValue(), func, ((PCPath) getValue()).getFieldMapping(state).getColumns()[0]);
        } else {
            func = dict.getCastFunction(getValue(), func);
{{        }}}

        int idx = func.indexOf("{0}");
        buf.append(func.substring(0, idx));
        getValue().appendTo(sel, ctx, state, buf, index);
        buf.append(func.substring(idx + 3));
{{    }}}

The part of code that generate the exception is:

        Path<String> authorPath = from.join("references").join("authors");
        predicateList.add(cb.like(cb.lower(authorPath), cb.literal('%' + value + '%')));

If I change this code to use another string function like trim(), it works (ie. the SQL query is correctly generated).

 

Regards,

Patrice

ps: remark that the only uses of DBDictionary.getCastFunction(Val val, String func, Column col) are for those classes.