C# 4 and .NET 4

Persisting Dataset Changes ❘ 853
New row updated and new RegionID assigned by database
1 Eastern Unchanged
2 Western Unchanged
3 Northern Unchanged
4 Southern Unchanged
5 North West Unchanged
Look at the last row in the DataTable. The RegionID had been set in code to 999, but after executing
the RegionInsert stored procedure, the value has been changed to 5. This is intentional — the
database will often generate primary keys for you, and the updated data in the DataTable appears
because the SqlCommand definition within the source code has the UpdatedRowSource property set to
UpdateRowSource.OutputParameters:
SqlCommand aCommand = new SqlCommand("RegionInsert", conn);
aCommand.CommandType = CommandType.StoredProcedure;
aCommand.Parameters.Add(new SqlParameter("@RegionDescription",
SqlDbType.NChar,
50,
"RegionDescription"));
aCommand.Parameters.Add(new SqlParameter("@RegionID",
SqlDbType.Int,
0,
ParameterDirection.Output,
false,
0,
0,
"RegionID", // Defines the SOURCE column
DataRowVersion.Default,
null));
aCommand.UpdatedRowSource = UpdateRowSource.OutputParameters;
code download DataAdapter2.cs
What this means is that whenever a data adapter issues this command, the output parameters should be
mapped to the source of the row, which in this instance was a row in a DataTable. The flag states what data
should be updated — the stored procedure has an output parameter that is mapped to the DataRow. The
column it applies to is RegionID because this is defined within the command definition.
The following table shows the values for UpdateRowSource.
uPdaTeroWsourCe Value
Both
FirstReturnedRecord
None
OutputParameters
desCriPTion
A stored procedure might return output parameters and also a complete
database record. Both of these data sources are used to update the
source row.
This implies that the command returns a single record, and that the contents of
that record should be merged into the original source DataRow. This is useful
where a given table has a number of default (or computed) columns because
after an INSERT statement these need to be synchronized with the DataRow
on the client. An example might be ‘INSERT (columns) INTO (table) WITH
(primarykey)’, then ‘SELECT (columns) FROM (table) WHERE (primarykey)’. The
returned record would then be merged into the original row.
All data returned from the command is discarded.
Any output parameters from the command are mapped onto the appropriate
column(s) in the DataRow.
www.it-ebooks.info