This is a Date Index. If you care about time, then turn around. If you care about timezones, turn around. If however you only wish for the utter humiliating destruction of timezones, then you have come to the right place.

The best way to describe this is to think about the use case(cover your eyes if you don't like dirty words): reports.

The core case for this is events/meetings/appointments/etc. Sure you care about time if you are displaying details but when reporting its about as useful as a monkey on a pogo stick. Here you want to do queries such as "show me how many appointments I had on May 08 2011." The query for that with the current DateTime catalog looks something like:

...
{'query':DateTime('2011-05-08',
         DateTime('2011-05-08')+.999999999),
  'range': 'min:max'}},
...

Without the .99999 a bunch of stuff gets filtered. Oh, and if your portal is in a different timezone than the server running it, forget about getting anything consistent.

Furthermore, we can send in just about any data you want and convert it. You like strings? We got strings. You like datetime? We got datetime. Oh, and it takes care of that pesky "/" vs "-" thing as well when parsing strings.

Of course, with careful code planning an monitoring we can avoid all this but who has time for that. Enter DateDateIndex. DateDateIndex is the counter to DateTimeIndex, which coincidentally was renamed to DateIndex, although it is actually a DateTime Index.

This index will take a date, time, datetime, DateTime, or string. It will discard any time and timezone information faster than you can say uncle. Then it will turn it into a time neutral DateTime object and store like any other. The difference being that you can query and not have to worry about timezones, since it will always be native to the local server.

To use, simply add a new DateDateIndex from the ZMI like you would any other. When you query, you can also query like you would, except its very flexible. The scenario above would look more like:

...
{'query':'2011-05-08',
         '2011-05-08'),
  'range': 'min:max'}},
...

If you send in times for the query they will get stripped too. Query can also take any meaningful data type like DateTime, datetime, etc...

Note that the current version has only been tested on Plone 4.x. Who knows what lies in store for the rest.

• Has this been extensively tested? Nope.
• Are there better ways to do this? Probably, but this seemed the fastest way at the moment.
• What is the performance like? No idea. Although this inherits from DateIndex so all those

optimizations will trickle down.

• Is this a good solution for me? Probably not. This is currently being tested before hitting a live site and is without issues but tread carefully. It's not SO insance.
• Only tested in Plone 4.0 and 4.1

https://github.com/eleddy/PluginIndexes.DateDateIndex