Tips & Tricks - Data from Views/Tables into multiple LUs
We use the procedure MIGRATE_SOURCE_DATA for these kind of jobs. Data
Migration offers a very flexible way of specifying both dataset to be selected as
well as methods to be executed and configuration of these. To understand the basic
principles behind execution of this procedure,
look here >>.
Contents
If you specify a View Name on insert of a new Header, this view name will also
be applied to the Method List folder and your first method will be inserted by default.
In the Formatting tab, Source Name may contain any Oracle data source; table/view/DB-link, JOINs and
even UNION. Look here >> for advanced
settings.
Source Owner is used only to specify another owner than Appowner. It is not recommended
to specify Appowner here, in case this job is exported to a database with another
name for Appowner.
Use these to build complete SQL-statements. Remember that this is SQL-syntax.
When entering data for a standard New__/Modify__, it is enough just to enter
Sequence and View Name.
On INSERT, Data Migration will perform following actions :
- Check that there is a Package with the name <view_name>_API, containig New__/Modify__
methods.
- Lookup dictionary to find all columns for this View.
- Perform New__ with PREPARE option, to catch Business Logics default-values.
- Lookup LUs belonging to IIDs in order to Enumerate these, showing all alternatives.
- Generating data into Method List Attributes
(available from RMB)
- Generating data into folder Source Mapping
according to defined Rules
Important Rules :
No. in Screenshot |
Rule |
Description |
1 |
ADDOBJID |
Create extra column detail for OBJID.
Creates a column at the end of Source Mapping.
This column enables us to Restart jobs |
2 |
CREATEDET |
Create new column details based on views in Method List.
If activated, rows will be inserted into Source Mapping. |
3 |
MAPDBCOL |
Remap IIDs client-default to DB-default.
If this rule is active, default-values for IID's with client-values will
be remapped to DB-values |
- |
SAVEJOBDAYS |
Saves info in history-tables, enables us to review previous
executions, including error lines.
It is recommended always to set this rule Active |
- Columns are fetched from the View Name in folder Method List.
- If PrefixOption is set to 'ViewName' , this will automatically prefix column
names. This is convenient if you have several methods in the method list, and
especially if there is lot of columns in the views, like in this example. Key-columns
are not prefixed.
- Source Column is automatically mapped if there is a match in Column Name
between Source Name and View Name
- Remember to map ROWID on last row. This will enable
Restart of jobs if errors occur.
The items in each methods Attribute-string are listed in Method List Attributes
(RMB on Method List).
- Only columns that have got a value from the Source Mapping dataset, will
appear on the attribute-string, and in the same order as Column Seq.
- If values are entered into Fixed Value, this value will override selected
value. No apostrophes in Fixed Value.
- The Modify-option will use columns with flag equal P or K to lookup the
view for Objid/Objversion to
perform method Modify__ with. These flags can be changed if you want to use
other columns for this lookup.
- Check Box for New and Modify are mapped according to Flags from the View-comments
('I' = OnNew, 'U'=OnModify).
Changing values for OnNew is not recommended unless special requirements in
Business Logic are present.
OnModify may be changed, but remember item 1 above.
- LU Reference show what references are connected to this column in ViewComments
:
- Master references for key columns (Flags = P or K)
- IID columns, alternate values are shown in column 'Iid Values'
- If Flags starts with M, this is a mandatory basic
data that has to be present in the reference LU before migration can start.
For details about error handling look here
>>