Why Does Java's SimpleDateFormat parse this

Question

Hi I\'ve got a simple date format set up with a custom format string: MMddyy

and I give it the following value to parse: 4 1 01

I don\'t think it should pars

Answer 1

This is expected behaviour - you are telling the DateFormat object to expect a 6 character String representation of a date and that is what you passed in. Spaces are parsed OK. However, if you used "4x1x01" you would get an error. Note that when parsing, leniency defaults to true e.g.

DateFormat df = new SimpleDateFormat("MMddyy");
Date date = df.parse("4 1 01"); // runs successfully (as you know)

DateFormat df = new SimpleDateFormat("MMddyy");
Date date = df.parse("41 01"); // 5 character String - runs successfully

DateFormat df = new SimpleDateFormat("MMddyy");
df.setLenient(false);
Date date = df.parse("41 01"); // 5 character String - causes exception

DateFormat df = new SimpleDateFormat("MMddyy");
Date date = df.parse("999999"); // 6 character String - runs successfully

DateFormat df = new SimpleDateFormat("MMddyy");
df.setLenient(false);
Date date = df.parse("999999"); // 6 character String - causes exception

When leniency is set to true (the default behaviour), the parse makes an effort to decipher invalid input e.g. the 35th day of a 31 day month becomes the 4th day of the next month.

Answer 2

for parsing the size of a pattern (number of repeated characters) is not the expected size of the corresponding text. From the javadoc, for the different relevant presentation types:

• Number: For parsing, the number of pattern letters is ignored unless it's needed to separate two adjacent fields.
• Year: During parsing, only strings consisting of exactly two digits […] will be parsed into the default century. Any other numeric string, such as a one digit string, a three or more digit string, or a two digit string that isn't all digits (for example, "-1"), is interpreted literally. So "01/02/3" or "01/02/003" are parsed, using the same pattern
• Month: If the number of pattern letters is 3 or more, the month is interpreted as text; otherwise, it is interpreted as a number.

The whitespace causes the parser to stop parsing the actual field (trailing spaces are not valid for numbers) and start with the next one. Since the pattern does not have a space between these two fields, it is not consumed and is part of the second field (leading spaces are valid). So the year got is not "exactly two digits" and will not be parsed into the default century.

Parsing tests (lenient set to false):

FORMAT   TEXT     RESULT (ISO yyyy-MM-dd)
-------------------------------------------------
ffffdyy    01011    2011-01-10  
ffffdyy    10 11    0011-01-10  (year is 3 chars: " 11")
ffffdyy    10 1     0001-01-10  (year is 2 char but not 2 digits: " 1")

ffffdy     01011    2011-01-10  ("y" same as "yy")

dd yy    10 11    2011-01-10  (ok, whitespace is consumed, year: "11")

d/y      3/4      0004-01-03  (year is not 2 digits)
d/y      3/04     2004-01-03  

M/d/y    4/6/11   2011-04-06

Answer 3

The 2 digit year is ambiguous - and it is therefore assuming 0001 - the first year that would have ended in 01. Can you convert to 4 digit years - maybe using String manipulation?