You want to import very large OpenDocument tables and worry about memory usage? In that case the opendocument-reader might be interesting for you. Instead of loading the entire document, it is a pull-based reader.
Here is an example how to use the opendocument-reader:
Document doc = new Document(new ZipFile("myTable.ods"));
Table table = doc.nextTable();
Row row = table.nextRow();
while( row != null ) {
Cell a = row.nextCell();
Cell b = row.nextCell();
System.out.println( a.getContent + "\t" + b.getContent );
row = table.nextRow();
}