OpenEuphoria

Re: structure

• Posted by gwalters at sc.rr.com Jul 26, 2001
• 357 views

interesting...i'll think on it some..

..george

----- Original Message -----
From: <irvm at ellijay.com>
To: "EUforum" <EUforum at topica.com>
Subject: Re: structure


>
>
> On Thursday 26 July 2001 09:35, gwalters at sc.rr.com wrote:
>
> > irv, you've hit it. each one can have a different length...problem is
fixed
> > now..thanks all for your input..
>
> That, of course, is one of Euphoria's advantages - you can pack
> pretty much anything you want into a sequence, without any arbitrary
> limits or having to pre-define the space.
>
> As an example, the 'normal' way to implement an A/R system is with
> two files: a customer file containing name, addr, balances, etc.
> and a transaction file, with current txns.
>
> With Euphoria, I tend to use just one file, which contains:
> (hope my indentation makes it thru EuForum intact)
>
> customer key
>   -- customer info
>        -- name
>        -- addr
>        -- balances
>             -- current
>             -- 30
>             -- 60
>             -- 90
>   -- customer txns (zero or more)
>         -- inv #
>         -- date
>         -- amt...
>
> As customer makes purchases, the new purchase is just appended to
> his/her txns sequence.
>
> Number of transactions can be obtained by calling length(customer[txns])
> Customer totals are computed by looping 1... length(txns), etc.
>
> This eliminates a lot of lookups and searching, since everything related
> is located within the one file. It also simplifies things such as deleting
> a customer account:
> if length(txns) > 0
> or balance != 0 then
>   error("Can't delete active account!")
>
> Backups are simple - one file to copy, as are statements.
> Reviewing a customer's account is verrry quick, as well.
>
> EOM reports are less quick, because you have to loop thru all
> customers and consolidate the data. Usually you're printing these
> reports, so there's plenty of time for processing while the printer
> chugs along.
>
> Starting a new period is just a matter of adjusting the 30/60/90,
> and then setting customer[x][txns] = {}
>
> You could keep the entire sequence of customers 'in-memory',
> but it's better to store these in a EDS database, since EDS will
> handle the lookup logic for you and ensure the use of unique keys
> automatically.
>
> Regards,
> Irv
>
>
>
>
>