PM10351: JPA2UML TRANSFORMATION ISSUES

APAR status

• Closed as program error.

Error description

• The JPA to UML transform generates a UML model which causes
  various problems when using this model to do UML to JPA
  transform, including compile errors and runtime errors.
  
  A. @JoinColumn annotations are generated with blank
  referencedColumnName.
  E.g. @JoinColumn(name = 'XXX_XXX_XXX', referencedColumnName =
  '')
  
  B. The original JPA source code has @Column annotations in
  XXXXXPK.java at the top of the class, but these are not
  translated into ??Column???Id?? for primary key columns in the
  UML model. ??Column?? is missing.
  
  C. For entities with One-Many relationships there are collection
  type fields with duplicate names, and you get a compile error.
  
  
  D. For entities with One-Many relationships there are duplicated
  NamedQuery and you get a compile error.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED:                                              *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  (A)  During the transform the referenced column name is not
  checked for null or empty.
  
  (B) The transformation was assuming there is no stereotype
  in the PK file, so the rule that creates the properties was
  consuming and blocking all the other property rules to run
  (column is one of them).
  
  (C,D) there is one Entity (E1) that has bidirectional
  relationship to another three entities (E2, E3,  E4). E1 has
  three propertes e2, e3, e4 and each one of them has
  annotation @mappedby(e1). while all the other Entities has a
  property (e1) with the join column/table.
  the JPA to UML transformation was trying to match the
  properties to define the ends of the three associations and
  it was only checking by property name ( which is e1 in the
  three cases) which means the three association was created
  using (E1.e2, E2.e1) (E1.e2, E3.e1) , (E1.e2, E4.e1) which
  is not correct ( leads to more than one property of the same
  name)
  

Problem conclusion

• (A)  A simple check was added to determine if the referenced
  column name is null or empty, and if true do not add it.
  
  (B) Actually there is no need to consume as each rule is
  checking for the annotation and will not work if the
  annotation is not set so that has been corrected.
  
  (C,D) The matching was changed to make sure that the
  property with the same mapped name has the same type of the
  second Entity, so the generated model from the JPA to UML
  will be created without errors and then issues C and D will
  not happen.
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Fix information

• Fixed component name

  SW ARCHITECT WI

• Fixed component ID

  5724I7001

Applicable component levels

• R754 PSN

     UP