Hello all, I am wondering about the memory arrangement in C#.

If I am going to load, say, 20000 data rows from about 20 tables at once,
where will be the data stored in my machine? All save on RAM? Or a portion of
them is fetched to RAM for processing?

Thanks.

Kenny,

The datarows are loaded in memory, where that is depends on your memory
management, however you cannot influence that in your dataset.

(By the way, probably nobody will find this a good idea what you try to do.
There is a lot to tell why that is not a good idea, however start thinking
about a not wanted power down of the computer or any other reason it goes
down).

I hope this helps,

Cor

Cor, Thanks much for your help.

So, how about... I construct a dataset containing tables like what I've put
in my SQL server, then if I wanna view/edit info of one of the records, I
just load the datarows from all tables which are related to it?

There is no need to load all data. All I need to do is construct the entire
table structure in my dataset...

(Please help me on this. Coz I am new to C# programming, and this is the
very first dataset programs I've ever dealt with.)

Thanks much.

"Cor Ligthert [MVP]" wrote:

> Kenny,
>
> The datarows are loaded in memory, where that is depends on your memory
> management, however you cannot influence that in your dataset.
>
> (By the way, probably nobody will find this a good idea what you try to do.
> There is a lot to tell why that is not a good idea, however start thinking
> about a not wanted power down of the computer or any other reason it goes
> down).
>
> I hope this helps,
>
> Cor
>
>
>

Kenny,

Try to get only that what is needed by the client in your dataset.
And do needed updates as soon as possible.

However, there is impossible to give a "best" solution. In a webform you
will need another solution than on a PDA and/or on a windowsform.

In all circumstances is a dataset disconnected by the way, which means that
you are self responsable to do the updates.

What everybody almost forever forget is that as soon as you have filled the
dataset in the computer, somebody else can have changed that again. If the
filled data is few, than the change on that is of course much less.

This concept is complete different from the recordset approach.

I hope this gives some ideas

Cor

"Kenny" <(E-Mail Removed)> schreef in bericht
news:A4863246-3ABE-4F65-A731-(E-Mail Removed)...
> Cor, Thanks much for your help.
>
> So, how about... I construct a dataset containing tables like what I've
> put
> in my SQL server, then if I wanna view/edit info of one of the records, I
> just load the datarows from all tables which are related to it?
>
> There is no need to load all data. All I need to do is construct the
> entire
> table structure in my dataset...
>
> (Please help me on this. Coz I am new to C# programming, and this is the
> very first dataset programs I've ever dealt with.)
>
> Thanks much.
>
> "Cor Ligthert [MVP]" wrote:
>
>> Kenny,
>>
>> The datarows are loaded in memory, where that is depends on your memory
>> management, however you cannot influence that in your dataset.
>>
>> (By the way, probably nobody will find this a good idea what you try to
>> do.
>> There is a lot to tell why that is not a good idea, however start
>> thinking
>> about a not wanted power down of the computer or any other reason it goes
>> down).
>>
>> I hope this helps,
>>
>> Cor
>>
>>
>>

Thanks much.

"Cor Ligthert [MVP]" wrote:

> Kenny,
>
> Try to get only that what is needed by the client in your dataset.
> And do needed updates as soon as possible.
>
> However, there is impossible to give a "best" solution. In a webform you
> will need another solution than on a PDA and/or on a windowsform.
>
> In all circumstances is a dataset disconnected by the way, which means that
> you are self responsable to do the updates.
>
> What everybody almost forever forget is that as soon as you have filled the
> dataset in the computer, somebody else can have changed that again. If the
> filled data is few, than the change on that is of course much less.
>
> This concept is complete different from the recordset approach.
>
> I hope this gives some ideas
>
> Cor
>