help with sparse matrix

So, if there is a "value" at the "coordinate", you create a node. That node will hold a "key" (the column) and a "value" ( the value stored in the column),

An easy implementation will use ArrayLists - one of which [ a "RowNode" ArrayList] will have as many items as the matrix has rows, and each entry in the ArrayList will hold as its "value" a reference to it's "tail" or "child" or "EdgeNode" array list, which will hold as many nodes as there are values in that row.

To perform your functions, you select the RowNode which reprents your row of interest, then walk that node's EdgeNode ArrayList. For each EdgeNode you have a column (the key) and a value. Extract that data and perform your function.

If you're going to do this, there is work involved here coding your RowNode and EdgeNode classes. [if you just used lists which would store "0" as a value for the columns where there is no value it would be easier ...] The most important part is gaining access to the EdgeList contained in your RowNode, and then reporting out the value held in each of the EdgeNodes.

Some psuedocode:

ArrayList<RowNode> rowList = new ArrayList<RowNode>();

create your RowNode arraylist.

for (int i = 0; i < #rows; i++ )
{
RowNode tempNode = new RowNode( new ArrayList<EdgeNode>() );
rowList.add( tempNode );
}

Now, for each column with a value in row 1, add an EdgeNode to the EdgeList held in (referenced by) the RowNode ...

for (row 0 to n-1 )
{
for (column 0 to m-1)
{
if( value )
{
create EdgeNode( key = column number, value = value in quadrant );
add this EdgeNode to the EdgeList - RowList.get(row number).add(EdgeNode);
}
}
}

take out the <> content ...