Optional parts in SimpleDateFormat

Question

I\'m reading in date strings that could be with or without a time zone adjustment: yyyyMMddHHmmssz or yyyyMMddHHmmss. When a string is missing a z

Answer 1

I have solved a similar problem some time ago by extending SimpleDateFormat. Below an crude implementation to show the idea of my solution. It may not be fully complete/optimised.

public class MySimpleDateFormat extends SimpleDateFormat {
    private static final long serialVersionUID = 1L;
    private static String FORMAT = "        
    private static int FORMAT_LEN = "yyyyMMddHHmmss".length();
    private static String TZ_ID = "GMT";

    public MySimpleDateFormat() {
            this(TimeZone.getTimeZone(TZ_ID));
    }

    public MySimpleDateFormat(TimeZone tz) {
            super(FORMAT);
            setTimeZone(tz);
    }

    @Override
    public Date parse(String source, ParsePosition pos) {
            // TODO: args validation
            int offset = pos.getIndex() + FORMAT_LEN;
            Date result;
            if (offset < source.length()) {
                    // there maybe is a timezone
                    result = super.parse(source, pos);
                    if (result != null) {
                            return result;
                    }
                    if (pos.getErrorIndex() >= offset) {
                            // there isn't a TZ after all
                            String part0 = source.substring(0, offset);
                            String part1 = source.substring(offset);
                            ParsePosition anotherPos = new ParsePosition(pos.getIndex());
                            result = super.parse(part0 + TZ_ID + part1, anotherPos);
                            if(result == null) {
                                    pos.setErrorIndex(anotherPos.getErrorIndex());
                            } else {
                                    // check SimpleDateFormat#parse javadoc to implement correctly the pos updates
                                    pos.setErrorIndex(-1);
                                    pos.setIndex(offset);
                            }
                            return result;
                    }
                    // there's something wrong with the first FORMAT_LEN chars
                    return null;
            }
            result = super.parse(source + TZ_ID, pos);
            if(result != null) {
                    pos.setIndex(pos.getIndex() - TZ_ID.length());
            }
            return result;
    }

    public static void main(String [] args) {
            ParsePosition pos = new ParsePosition(0);
            MySimpleDateFormat mySdf = new MySimpleDateFormat();
            System.out.println(mySdf.parse("20120622131415", pos) + " -- " + pos);
            pos = new ParsePosition(0);
            System.out.println(mySdf.parse("20120622131415GMT", pos) + " -- " + pos);
            pos = new ParsePosition(0);
            System.out.println(mySdf.parse("20120622131415xxx", pos) + " -- " + pos);
            pos = new ParsePosition(0);
            System.out.println(mySdf.parse("20120x22131415xxx", pos) + " -- " + pos);
    }
}

The gist is that you need to check the input string and "guess" somehow that the TZ field is missing, add it if so and then let the SimpleDateFormat#parse(String, ParsePosition) do the rest. The implementation above isn't updating ParsePosition according to the contract in javadoc in SimpleDateFormat#parse(String, ParsePosition)

The class has a single default ctor as there's only one format allowed.

The method MySimpleDateFormat#parse(String, ParsePosition) is invoked by SimpleDateFormat#parse(String) so it's sufficient to cover both cases.

Running the main() this is the output (as expected)

Fri Jun 22 14:14:15 BST 2012 -- java.text.ParsePosition[index=14,errorIndex=-1]
Fri Jun 22 14:14:15 BST 2012 -- java.text.ParsePosition[index=17,errorIndex=-1]
Fri Jun 22 14:14:15 BST 2012 -- java.text.ParsePosition[index=14,errorIndex=-1]
null -- java.text.ParsePosition[index=0,errorIndex=5]