Development of own data originsIn the most cases the filtering is activated by FormattedField, which set FilteredTableSource and FilteredControlSource. For all components refering at the TableElement which has to be filtered, HSRecordSetCondition were created and in the RecordSet by FormattedField HSRecordSetCondition were created and set in setRecordSetConditions(RecordSetCondition[] conditions).
After a change of a filter value a refresh()
is disposed on the underlying RecordSet
by the TableElement.
The refresh() method of the RecordSets can read this HSRecordSetCondition`s
by getRecordSetConditions()
and confine the amount.
Usually the sorting is affected of the TableElement,
or defined by the eventually used SScrollTable.
At this for every sorting choice a RecordSetOrderCondition is generated and set
in the RecordSet by setRecordSetOrderConditions(RecordSetOrderCondition[] conditions).
These can be rode and edit for example in the refresh() method of RecordSet by
getRecordSetOrderConditions().
At the positioning the cursor is placed on another sentence inside the available amount of data. Cusor means the actual selected row which has to be readable by getRow(). This stands at the Initialisation before the first dataset. Following methods are addressed.
| Methode | Aufgabe |
|---|---|
| previos() | the cursor is moved one position down |
| next() | the cursor is moved one position up |
| first() | the cursor is moved at the first position |
| last() | the cursor is moved at the last position |
| absolute(int i) | the cursor is set at the commited row. |
| beforeFirst() | the cursor is set before the first data set |
| afterLast() | the cursor is set after the last data set |
| search(RecordSetCondition[] searchConditions) | Positions the data set on the basis of the commited RecordSetCondition`s. In this case it is no filtering, but it is searched after the conditions within the amount of the result and then the cursor is set at the first sentence. At this it is no refresh() called. Mostly the SearchedTableSource in the FormattedField is responsible for this. |
| moveToInsertRow() | Positions also after the last data set whereas inserted should be set to true. This will be explained in the next chapter. |
At the processing of data should be worked with a extra cache for the actual row for making changes undoable.When an Object is updated setRowUpdated() should be set to true.
By the call of next() the changes of the actual row will be proofed. Is the property UpdateConfirm in the TableElement activated then a dialog asks for the storage.
Is this accept then a updateRow() will be called and the data of the cache have to be taken in the original stock.
The function insertRow() is called when isInserted() returns true. These has to be set by the user in moveToInsertRow(). moveToInsertRow() is called by eva/3 when it wants to create a new data set. isInserted() has to return true after the call of moveToInsertRow(). These can't return false until a data set is saved or the cursor has a new position.In the insert Mode (isInserted()) you should be attend to that getRowCount() returns one data set more.
| Methode | Aufgabe |
|---|---|
| getObject(int columnIndex) | Returns the Object for the comitted column of the actual row. |
| updateObject(int columnIndex,Object x) | Sets the Object for the commited column of the actual row. Here
should be proved if the commited Object differs from the set Object and then set setRowUpdated(true). |
| updateRow() | Saves the changes of the actual row if isUpdated() is true |
| insertRow() | Adds a new row if isUpdated() and isInserted() is true. |
| moveToInsertRow() | Increases the row count temporary about 1 und positions at the last new row. Furthermore isInserted should return true. |
| cancelRowUpdates() | Aborts the actual changes, sets updated and inserted to false and clears the update cache. Here is the last data set positioned. |
Quod vide CSVRecordSource reference implementation.