Accumulative sorting in ADF 10g tables

Just what the title says.

In an application there were a few tables filled with results of some queries. I haven’t yet figured out why, but it was of a high importance to the client that all the tables had accumulative sorting (like the order by clause in SQL). This requirement took me a while to figure out how to implement, whilst the solution was quite elegant and pretty small to write it here.

So, there’s a table with results of a search (the query really doesn't matter as the entire sorting solution will be implemented in the view layer)

<table value="#{bindings.CustomerSearch.collectionModel}" var="row"

rows="#{bindings.CustomerSearch.rangeSize}"

first="#{bindings.CustomerSearch.rangeStart}" banding="row"

bandinginterval="1"

emptytext="#{bindings.CustomerSearch.viewable ? 'No rows yet.' : 'Access denied.'}"

sortlistener="#{CustomersBean.onSort}">

<column sortproperty="Customerid" sortable="true"

headertext="#{bindings.CustomerSearch.labels.Customerid}">

<commandlink text="#{row.Customerid}"

onclick="return initiateReplacement(#{row.Customerid});"

shortdesc="Click to initiate customer replacement"/>

</column>

<column sortproperty="Surname" sortable="true"

headertext="#{bindings.CustomerSearch.labels.Surname}">

<outputtext value="#{row.Surname}"/>

</column>

<column sortproperty="Name" sortable="true"

headertext="#{bindings.CustomerSearch.labels.Name}">

<outputtext value="#{row.Name}"/>

</column>

<column sortproperty="Fathername" sortable="true"

headertext="#{bindings.CustomerSearch.labels.Fathername}">

<outputtext value="#{row.Fathername}"/>

</column>

<column sortproperty="Address" sortable="true"

headertext="#{bindings.CustomerSearch.labels.Address}">

<outputtext value="#{row.Address}"/>

</column>

</table>

view raw gistfile1.xml hosted with ❤ by GitHub

The whole trick is to set sortListener=”#{CustomersBean.onSort}” for the table. The backing code for this sortListener is found in the backing bean named CustomerBean.

	import oracle.adf.view.faces.event.SortEvent;
	import oracle.adf.view.faces.model.SortCriterion;
	import oracle.adf.view.faces.component.core.data.CoreTable;
	import java.util.List;
	import java.util.ArrayList;
	public class Customers {
	public void onSort(SortEvent se) {
	// get selected criterion
	List sortList = se.getSortCriteria();
	SortCriterion currentSortCriterion = (SortCriterion)sortList.get(0);
	// get previously selected criteria
	ArrayList persistentSortList = getCustomerTableSortCriteria();
	// the rest of the code expands the sort criteria, applies the new criteria list to the table and saves the criteria list for the future
	ArrayList updatedSortList = new ArrayList();
	updatedSortList.add(0, currentSortCriterion);
	for(Object o : persistentSortList) {
	updatedSortList.add((SortCriterion)o);
	}
	CoreTable ct = (CoreTable)se.getComponent();
	ct.setSortCriteria(updatedSortList);
	setCustomerTableSortCriteria(updatedSortList);
	}
	public static final String CUSTOMER_TABLE_SORT_CRITERIA = "customerTableSortCriteria";
	/* As the sort criteria returned at any call to the onSort method have request scope,
	* I use a session scoped bean that I call viewScope (I clean it up every time the viewId is changed, meaning that user has entered another page) which in fact is a java.util.HashMap.
	* The method getCustomerTableSortCriteria() returns the Object under key CUSTOMER_TABLE_SORT_CRITERIA stored in viewScope.
	* Naturally instead of viewScope one can use any other session scoped bean to store the ArrayList of SortCriterion objects.
	*/
	private ArrayList getCustomerTableSortCriteria() {
	ArrayList sortCriteria = (ArrayList)viewScope.get(CUSTOMER_TABLE_SORT_CRITERIA);
	if(sortCriteria == null) {
	sortCriteria = new ArrayList();
	setCustomerTableSortCriteria(sortCriteria);
	}
	return sortCriteria;
	}
	private void setCustomerTableSortCriteria(ArrayList sortCriteria) {
	viewScope.put(CUSTOMER_TABLE_SORT_CRITERIA, sortCriteria);
	}
	}

view raw gistfile1.java hosted with ❤ by GitHub

And that’s pretty much it. The table is now sortable accumulatively: you can hit customerid first, then name, then surname etc. the resulting table will look similarly to what produces SQL clause order by customerid, name, surname, etc.
Hope this helps.