Arrays instead of Sql

[Phil Davis]

Sarah Reichelt wrote:
> On Tue, May 19, 2009 at 10:30 AM, Tom Cole <tom.cole at asu.edu> wrote:
>   
>> I've heard that instead of using a backend like sql to manage a database, I
>> could stay in Revolution and use arrays. I know nothing of arrays.
>>
>> I've written here before about a bird database I have. Several times.
>> Forgive.
>>
>> I just have a few items in a record:
>> birdname,birdnotes,place,placenotes,date  However, there are about 18000
>> birds (records) and I've got them all in one field in Rev now and it's too
>> slow to sort.
>>     
>
> Just as a first test, what if you put the data into a custom property
> instead of a field?
> Use this CP as the data storage and grab what you need as you need it. e.g.
>
> -------------------------------------------------------------
> put the cDataStore of this stack into tData   -- grab the complete
> data set out of the custom property into a variable
> filter tData with "*" & fld "Filter" & "*"   -- assuming you want to
> extract a sub-set
> set the itemDel to tab  -- or whatever delimits your columns
> sort lines of tData by item 2 of each   -- choose the column to sort by
>
> put tData into fld "Display"  -- show the final results
> -------------------------------------------------------------
>
> Even if you end up with the complete set of data after the filter,
> sorting a variable is enormously faster than sorting a field.
> And getting data out of or into a custom property is again, greatly
> faster than for a field.
> I have done this successfully with multiple CPs each holding about
> 40,000 records and operations only take miliseconds.
>
> Cheers,
> Sarah

Paul Looney's "Ahsomme" multi-user business system is built on this CP 
approach to data storage, and its speed is quite acceptable - and all 
the data stacks are on a network file server! All the system's "brains" 
are in the client apps.

-- 
Phil Davis

PDS Labs
Professional Software Development
http://pdslabs.net