FWD » Core Development » Database

I have the following test case:

DEFINE TEMP-TABLE Main NO-UNDO
    FIELD Main_id  AS RECID XML-NODE-TYPE "HIDDEN":U.

DEFINE TEMP-TABLE Second NO-UNDO NAMESPACE-URI "http://www.openapplications.org/oagis/10":U
    FIELD Main_id AS RECID XML-NODE-TYPE "HIDDEN":U.

DEFINE TEMP-TABLE Third NO-UNDO NAMESPACE-URI "http://www.openapplications.org/oagis/10":U
    FIELD Main_id_id AS RECID XML-NODE-TYPE "HIDDEN":U.

DEFINE DATASET dsProcessField  XML-NODE-TYPE "HIDDEN":U
    FOR Main, Second, Third
    PARENT-ID-RELATION FOR Main, Second PARENT-ID-FIELD Main_id
    PARENT-ID-RELATION FOR Main, Third  PARENT-ID-FIELD Main_id.

create Main.

create Third.
Third.Main_i = recid(Main).

release Third.
release Main.

In OE this code does compile, and correct relationship field is picked Main_id_id.

In FWD, the wrong PARENT-ID-FIELD is picked at conversion time:

final DataSet dsProcessField = DataSetManager.define("dsProcessField", false, false)
         .setDefiningType(Start.class)
         .setXmlNodeType("HIDDEN")
         .forBuffers(main, second, third)
         .addParentIdRelation("ParentRelation1", main, second, "main_id")
         .addParentIdRelation("ParentRelation2", main, third, "main_id")
         .generate();

I guess the real problem here is that while parsing PARENT-ID-FIELD Main_id phrase we try to infer the the table used in the child relationship just with field name. And in this case the abbreviation name matches the name of the Second field name.

I think to solve the problem we need to set the currentRecord while parsing PARENT-ID-RELATION FOR phrase in for_parent_relation_spec.

I have put a conversion with 10611a of 9457 application, and manage to create a small test case.

def temp-table tt1 field f1 as char
                   field f2 as char
                   field f3 as char.

def temp-table tt2 field f3_abc as char
                   field f3_asd as char
                   field tt1_recid as recid.

define dataset dtt for tt1, tt2
       PARENT-ID-RELATION FOR tt1, tt2 PARENT-ID-FIELD tt1_recid
                                       PARENT-FIELDS-BEFORE (f1, f2, f3).

I have looked over the Progress documentation:

PARENT-ID-RELATION FOR parent-buffer-name, child-buffer-name

PARENT-FIELDS-BEFORE ( parent-field1[ , parent-field2]... )
    An optional phrase that specifies the fields from the parent-buffer-name temp-table record that are to be written to XML or JSON before any child-buffer-name temp-table records.

According to documentation, PARENT-FIELDS-BEFORE statement should included only fields from parent-buffer-name buffer (tt1 in our case).

The error that I get is:

'f3' is ambiguous; 2 matches found [tt2.f3_asd, tt2.f3_abc]
     [java]     at com.goldencode.p2j.schema.Namespace.findUnambiguously(Namespace.java:656)
     [java]     at com.goldencode.p2j.schema.Namespace.find(Namespace.java:528)
     [java]     at com.goldencode.p2j.schema.Namespace.find(Namespace.java:465)
     [java]     at com.goldencode.p2j.schema.SchemaDictionary.findFieldNode(SchemaDictionary.java:4603)

In this case FWD looks for field name in child-buffer-name buffer which is incorrect according to the docs.

So I think that 10611a exposed a existent problem in FWD, I will look into that.

10611a, rev. 16181 has a fix for #10611-8, I have converted #9457 application and it now passes.

def temp-table tt1 field f1 as int.
                   //field b as int.

def temp-table tt2 field b1 as int
                   field b2 as int.

def var h as handle.

create query h.
h:SET-BUFFERS (buffer tt1:handle, buffer tt2:handle).

h:QUERY-PREPARE ("for each tt1 where f1 = 10, first tt2 where b = 5").
h:query-open().

I have found a case in customer application, where a search for abbreviated field name will cause an error. Search for b field name for tt2 buffer will raise an ambiguous error (it can't decide between b1 and b2 field names). I think raising of the error should be postponed, until all scopes of the query has been searched.

Committed on 10611a, rev. 16182.