为什么解析此日期字符串会引发不可解析的日期异常？

Question

I'm using SimpleDateFormat with the pattern EEE MM/dd hh:mma , passing in the date String Thu 10/9 08:15PM and it's throwing an Unparseable date exception. Why? I've used various patterns with SimpleDateFormat before so I'm fairly familiar with its usage. Maybe I'm missing something obvious from staring at it too long.

The other possibility is funky (technical term) whitespace. The context is a screen-scraping app, where I'm using HtmlCleaner to tidy up the messy html. While I've found HtmlCleaner to be pretty good overall, I've noticed strange issues with characters that look like whitespace but aren't recognized as such with a StringTokenizer, for example. I've mostly worked around it and haven't dug into the character encoding or anything like that but am starting to wonder.

Answer 1

To test if it's the date format, write a test class to prove it out. For these types of things, I like to use bsh (beanshell). Here was my test:

sdf = new java.text.SimpleDateFormat("EEE MM/dd hh:mma");
System.out.println(sdf.format(sdf.parse("Thu 10/9 08:15PM")));

Which outputted: Fri 10/09 08:15PM

So, at least with my jdk / jre version (1.6), the format strings seem to work just fine. i think the next step is to make sure the string you're dealing with is exactly what you think it is. Can you add logging to your code, and dump out the input string to a log file? Then you could look at it in a nice text editor, run it through your test class, or look at it in a hex editor to make sure that it's just normal text.

Good luck!

Answer 2

First question here on StackOverFlow so I'm not sure what the proper way to mark this resolved is. Most of the answers are in the comments of Eric's answer.

The root cause was a 'space' character in the date string that was not recognized as such. It was a hex char of 'A0', which is a non-breaking space. I ended up converting the date string to a char array, checking the characters with Character.isSpaceChar(), and replacing those that returned true with a " " char.

Answer 3

Try this instead for your pattern:

EEE MM/d hh:mma

The difference is the single d instead of double dd , since your date is for 10/9 instead of 10/09.