If you have a lot of data and local arrays become too slow, you could use a global in cachetemp indexed by $J so that each "user" has their own space to work in. If you don't want to create a global mapping, all globals who's name begins with "CacheTemp" are mapped to the CACHETEMP database. Cache' SQL uses ^CacheTemp so you might want to use ^CacheTempX (for example).
"Max Sebastiani" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED] > Cheers Ram�n, and > thanks to everybody for your replies! > > Wow, these are great news for me: > Cache' seems the best technology for this kind of things. > > Performance is not an issue in our case, but using sparse arrays improves > the quality of code a lot, and this makes the difference for us, since we're > going to re-sell the solution to more than one customer. > > Max > > > "Max Sebastiani" <[EMAIL PROTECTED]> ha scritto nel messaggio > news:[EMAIL PROTECTED] > > Hello, I'm wondering if sparse arrays in COS are suitable for > > multidimensional data analysis. > > > > My intention is to store information into arrays like: > > > > s ProductsSold("Apple","1997","december","Venezuela")=2000 > > > > and create some routines to manipulate them (collapse dimensions, apply > > functions to each term, distribute values, etc). > > > > My first question is: are sparse arrays (in memory, not globals) limited > to > > 32k ? > > > > Max > > > > > >
