«

I am trying to get my head around dataset merging but despite a little
research I could still use a pointer (or ten). Basically I want to perform
an outer join operation on a dataset (created from a file operation) and the
resulting sql query based on the contents of said dataset. Effectively I
would end up with two dataset tables in my application. If I were writing
this to action from two tables it would be thus:



SELECT ds.id, db.*
from
dataset ds LEFT OUTER JOIN
database db on db.field1 = ds.id

such that I retain the original dataset items. I have read something about a
dataset relationship but I do not see how this achieves the all-inclusive
results I am after.

any insight much appreciated

n

The dataset "merge" method can do about 2 things from my experience.

1. If you have N number of different datatables, it can merge the N number
of tables into 1 (common) dataset.

2. If you have different ROWS, but the same table, you can merge the rows
together. Most times, the rows will be different.

3. A combo of 1 and 2.

If you try to merge rows and the rows have the same primary key, but you're
trying to merge columns of the same row, you'll have trouble.

Back to your question. You can't do what you're trying with a dataset. the
..select and .merge methods are very very basic.

You need to look at LINQ. Or you'll have to loop and do your logic if
you're in a pre 3.5 framework.

...



"Neil Chambers" <> wrote in message
news:%...
>I am trying to get my head around dataset merging but despite a little
>research I could still use a pointer (or ten). Basically I want to perform
>an outer join operation on a dataset (created from a file operation) and
>the resulting sql query based on the contents of said dataset. Effectively
>I would end up with two dataset tables in my application. If I were writing
>this to action from two tables it would be thus:
>
>
>
> SELECT ds.id, db.*
> from
> dataset ds LEFT OUTER JOIN
> database db on db.field1 = ds.id
>
> such that I retain the original dataset items. I have read something about
> a dataset relationship but I do not see how this achieves the
> all-inclusive results I am after.
>
> any insight much appreciated
>
> n

Aha!

the primary key property does the trick. When I merge now it does so based
on the PK and I get the exact inclusive results I am after!

Many, many thanks!

: n)

"sloan" <> wrote in message
news:O6c3%...
>
> The dataset "merge" method can do about 2 things from my experience.
>
> 1. If you have N number of different datatables, it can merge the N
> number of tables into 1 (common) dataset.
>
> 2. If you have different ROWS, but the same table, you can merge the rows
> together. Most times, the rows will be different.
>
> 3. A combo of 1 and 2.
>
> If you try to merge rows and the rows have the same primary key, but
> you're trying to merge columns of the same row, you'll have trouble.
>
> Back to your question. You can't do what you're trying with a dataset.
> the .select and .merge methods are very very basic.
>
> You need to look at LINQ. Or you'll have to loop and do your logic if
> you're in a pre 3.5 framework.
>
> ..
>
>
>
> "Neil Chambers" <> wrote in message
> news:%...
>>I am trying to get my head around dataset merging but despite a little
>>research I could still use a pointer (or ten). Basically I want to perform
>>an outer join operation on a dataset (created from a file operation) and
>>the resulting sql query based on the contents of said dataset. Effectively
>>I would end up with two dataset tables in my application. If I were
>>writing this to action from two tables it would be thus:
>>
>>
>>
>> SELECT ds.id, db.*
>> from
>> dataset ds LEFT OUTER JOIN
>> database db on db.field1 = ds.id
>>
>> such that I retain the original dataset items. I have read something
>> about a dataset relationship but I do not see how this achieves the
>> all-inclusive results I am after.
>>
>> any insight much appreciated
>>
>> n
>
>