Today vs. Status(CurrentDate)

[cbsteven]

My database of orders placed in our store has a script which is supposed to only display orders which are open and were placed at least two days ago. To accomplish this I search by a record's OrderAge field. The field definition of this was:

DayofYear(Today) - DayofYear(Creation Date)

However like many people I found the long Today Calculation at startup very annoying. I did some search on this forum and it looks as though everyone recommends status(CurrentDate) as a superior replacement for Today.

I changed the OrderAge function to use Status(CurrentDate). At first glance it seemed to be behaving normally, but the next day I checked and the records all had the same age that they had the previous day. I went in to the field def and turned off indexing. This made the the age update, but when I performed the 2-day Search option over a network it was VERY slow (probably as slow as the normal today calculation)

Any better ideas on how to implement this without a long delay at startup or at execution?

[Vaughan]

No matter what, the unstored calculation has to be performed some time or other.

One solution may be to change how you perform the search: instead of doing the calculation in each record, do it int he find request. By that I mean this:

You want to search for records that are 2 days old. One way is to create a calc field that works out the age in each record -- this is slow because the calc is done once for each record. The other way is to perform the find for the date that is 2 days ago: in this case the calc is peformed once -- in the find request -- not in each record. Much faster.

The other reason for not using Today function is that it does not recalculate itself if the file is hosted on a server (at least not without a whole lot of mucking about). This I consider more important than the calculating at startup.