Heiko
Well-known member
Any experts on strongly typed datasets out there ?
We have a strongly typed dataset.
Its contents is being shown in a listview.
The details (single rows) are being shown in a seperate "details" section, both on the same usercontrol.
The user may or may not update some fields, delete rows or add rows.
The entire dataset is then given to a "DataChangeManager" class, that will perform special operations, will also write a history of changes and so on.
So far, so good.
Unfortunately, just to collect data from a single control on the usercontrol back to the dataset, we have to write x lines of code!
That is, of course, pretty much annoying.
But for now, the only alternative I see - with way better generilzation is to use a helper class
which would use the reflection API to adress the correct table and correct row in the DS.
However, this would mean giving up all the benefits of strongly typed datasets! No intellisense. No type checking.
Has anyone ever had a similar problem?
Tnx
Heiko
We have a strongly typed dataset.
Its contents is being shown in a listview.
The details (single rows) are being shown in a seperate "details" section, both on the same usercontrol.
The user may or may not update some fields, delete rows or add rows.
The entire dataset is then given to a "DataChangeManager" class, that will perform special operations, will also write a history of changes and so on.
So far, so good.
Unfortunately, just to collect data from a single control on the usercontrol back to the dataset, we have to write x lines of code!
Code:
if not DS.isField1Null then
if DS.Field1 <> control.Text then
if control.Text = "" then
DS.Field = dbNull
else
DS.Field = control.text
end if
end if
else
if contol.text <> "" then
DS.Field = control.text
end if
end if
That is, of course, pretty much annoying.
But for now, the only alternative I see - with way better generilzation is to use a helper class
Code:
Sub Helper (aControl as control, aDS as Dataset, aTableName as string, aFieldName as String)
However, this would mean giving up all the benefits of strongly typed datasets! No intellisense. No type checking.
Has anyone ever had a similar problem?
Tnx
Heiko