• The Idea:
  • When working with numerical data one often needs to deal with missing values
• An Example:
  • People sometimes forget to enter weekly expenditures into their household budget program

• Given What You Know:
  • You would probably use double values and a sentinel like -1 for missing values
• Shortcomings:
  • In many situations there is no value that can be used as a sentinel
  • If you forget to check for the sentinel, the calculations will be wrong but the defect will be difficult to localize

• Ideally:
  • Every data type would have a sentinel
• Unfortunately:
  • The don't
• Fortunately:
  • null can be used as a sentinel for all reference types

• When "Collecting" the Data:
  • Declare a wrapper object (like Double or Integer)
  • If the information isn't missing, use the static valueOf() method to construct the wrapper object, otherwise leave it null
• When "Processing" the Data:
  • Determine if the wrapper object is null
  • If it is, treat it as missing. If it isn't use the doubleValue() or intValue() method to retrieve the value

javaexamples/programmingpatterns/MissingValues.java (Fragment: defaultmissing)

        total = 0.0;
        for (int i = 0; i < data.length; i++) {
            if (data[i] == null) {
                total += defaultValue; // Initialized elsewhere
            } else {
                total += data[i].doubleValue();
            }
        }
        average = total / (double) data.length;
        

javaexamples/programmingpatterns/MissingValues.java (Fragment: ignoremissing)

        total = 0.0;
        n     = 0;
        for (int i = 0; i < data.length; i++) {
            if (data[i] != null) {
                total += data[i].doubleValue();
                n++;
            }
        }
        average = total / (double) n;
        

javaexamples/programmingpatterns/MissingValues.java (Fragment: propagatemissing)

        missing = false;        
        total = 0.0;
        for (int i = 0; i < data.length; i++) {
            if (data[i] == null) {
                missing = true;
                break; // No reason to continue iterating
            } else {
                total += data[i].doubleValue();
            }
        }

        if (missing) {
            result = null;
        } else {
            result = Double.valueOf(total / (double) data.length);
        }