Wrong generated model source code

I have experienced some problem with LightSpeed, exatly in automatic code generation mechanism.

I have an some data structure in Oracle DB, and create LightSpeed data model based on it.

I DB exists transition table with only two columns wich is ForeingKey for other tables and makes up the PrimaryKey together. Special pecularity is both columns have underline in names. But LightSpeed created attributes in CamelCase form in accordingly entity without specifying any type of refference to original column name. As a result SQL script generated by LightSpeed is incorrect - DB dosn't have a column with name written in CamelCase form.

So, the next generated SQL code is wrong ("CategoryId, WorkflowId" should be named as "category_id, workflow_id"):

BEGIN

OPEN :cur1 FOR
SELECT
  CATEGORY_REL.CategoryId AS "category_0",
  CATEGORY_REL.WorkflowId AS "workflow_0"
FROM
  CATEGORY_REL
WHERE
  CATEGORY_REL.WorkflowId = 101;

END;

Thanks for drawing our attention to this.  We'll investigate why LightSpeed isn't assigning the underscored column names, but in the meantime the workaround is to manually enter the original Oracle column names (e.g. category_id) into the Column Name box.

This will be fixed in nightly builds dated 9 Sept 2010 and above, available from about 1500 GMT.  Thanks again for reporting this issue.

Thanks, i'm looking forward for fixed build.

And one more question, is it posible to purchase the LightSpeed other way than pay by Credit Card. In Russian Federation it is problem for organisation to pay by Credit Card.

Hi,

We accept other payment methods such as direct bank payment. If you would like to manage your purchase this way please get in touch: jd@mindscape.co.nz

I can invoice you or your organisation directly to complete the order.

I hope that helps,

John-Daniel Trask

Hello, Ivan.

Returning to our problem

[quote user="ivan"] but in the meantime the workaround is to manually enter the original Oracle column names (e.g. category_id) into the Column Name box.[/quote]

I have done as you had sugested earlier - I have entered the original name of column into Column Name property (in Properties window). Below is resulting SQL query to base, as you can see WHERE clause have changed, but "WorkflowId" column have still been presented in SELECT clause.

BEGIN

OPEN :cur1 FOR
SELECT
  CATEGORY_REL.CategoryId AS "category_0",
  CATEGORY_REL.workflow_id AS "workflow_0",
  CATEGORY_REL.WorkflowId AS "workflow_1"
FROM
  CATEGORY_REL
WHERE
  CATEGORY_REL.workflow_id = 123;

END;

Is it normal behavior of CodeGenerator?

LightSpeed Express build on 07 Sep 2010

[quote user="dtyger"]

Hello, Ivan.

Returning to our problem

[quote user="ivan"] but in the meantime the workaround is to manually enter the original Oracle column names (e.g. category_id) into the Column Name box.[/quote]

I have done as you had sugested earlier - I have entered the original name of column into Column Name property (in Properties window). Below is resulting SQL query to base, as you can see WHERE clause have changed, but "WorkflowId" column have still been presented in SELECT clause.

[/quote]

After reading you blog I have slightly chaged a model. I have specified a "Key Property Reference" for both associations: for Workflow to CategoryRel was "Id.WorkflowId", and for Category to CategoryRel was "Id.CategoryId".

After that generated SQL query looks good, but I'm worried about one thing: in CetegoryRel Entity properties "Identity Column Name" specified as "WORKFLOW_ID", but I expected, that here should be both of them workflow_id and category_id.

Below generated code:

[Serializable]
[System.CodeDom.Compiler.GeneratedCode("LightSpeedModelGenerator", "1.0.0.0")]
  public partial struct CategoryRelId
  {
    
    public CategoryRelId(
      int categoryId ,
      int workflowId     )
    {
      _categoryId = categoryId;
      _workflowId = workflowId;
    }
    
    #region Fields
 
    [Column("category_id")]
    private readonly int _categoryId;
    [Column("workflow_id")]
    private readonly int _workflowId;
 
    #endregion
    
    #region Properties
 
    public int CategoryId
    {
      get { return _categoryId; }
    }
    public int WorkflowId
    {
      get { return _workflowId; }
    }
 
    #endregion
    
    //... skipped
  }

[Serializable]
  [System.CodeDom.Compiler.GeneratedCode("LightSpeedModelGenerator", "1.0.0.0")]
  [System.ComponentModel.DataObject]
  [Table("CATEGORY_REL", IdColumnName="WORKFLOW_ID")]
  public partial class CategoryRel : Entity<CategoryRelId>
  {
    #region Relationships
 
    [ForeignKeyField("Id.CategoryId")]
    [ReverseAssociation("CategoryRels")]
    private readonly EntityHolder<Category> _category = new EntityHolder<Category>();
    [ForeignKeyField("Id.WorkflowId")]
    [ReverseAssociation("CategoryRels")]
    private readonly EntityHolder<Workflow> _workflow = new EntityHolder<Workflow>();
     #endregion
    
    #region Properties
 
    public Category Category
    {
      get { return Get(_category); }
      set { Set(_category, value); }
    }
 
    public Workflow Workflow
    {
      get { return Get(_workflow); }
      set { Set(_workflow, value); }
    }
    #endregion
  }

Identity Column Name is used for the implicit Id field that you get with a non-composite-keyed entity.  When the entity has a composite key, it is overridden by the per-key Column Name settings.  So in your case it will be ignored.