Question

I have been trying to use OleDbDataAdapter to update a DataTable but got confused about the commands.
Since I sometimes get info from diffrent tables I can’t use a CommandBuilder.
So I have tried to create the commands on my on but found it hard with the parameters.
DataTable.GetChanges returns rows that needs to use an INSERT or an UPDATE command – I guess I can’t distinct between them.
I need you to complete the following:

DataTable dt = new DataTable();
OleDbDataAdapter da = new OleDbDataAdapter();
// Here I create the SELECT command and pass the connection.
da.Fill(dt);
// Here I make changes (INSERT/UPDATE) to the DataTable (by a DataGridView).
da.UpdateCommand = new OleDbCommand("UPDATE TABLE_NAME SET (COL1, COL2, ...) VALUES (@newVal1, @newVal2, ...) WHERE id=@id"); // How can I use the values of the current row (that the da is updating) as the parameters (@newVal1, @newVal2, id....)?

Thank you very much!

Answer 1

The data adapter can work in conjunction with the datatable. As such, I’ve actually wrapped mine together into a class and works quite well. Aside from the complexities of my stuff, here’s a snippet that might help you along. When adding a parameter, you can identify the column source that the data is coming from FROM the DataTable. This way, when a record is internally identified as “Added” or “Updated” (or “Deleted”), when you build your SQL Insert/Update/Delete commands, it will pull the data from the columns from the respective rows.

For example. Say I have a DataTable, primary Key is “MyID” and has columns “ColX, ColY, ColZ”. I create my DataAdapter and build out my select, update, delete commands something like… (? is a place-holder for the parameters)

DataAdapter myAdapter = new DataAdapter()

myAdapter.SelectCommand = new OleDbCommand();
myAdapter.InsertCommand = new OleDbCommand();
myAdapter.UpdateCommand = new OleDbCommand();
myAdapter.DeleteCommand = new OleDbCommand();

myAdapter.SelectCommand.CommandText = "select * from MyTable where MyID = ?";
myAdapter.InsertCommand.CommandText = "insert into MyTable ( ColX, ColY, ColZ ) values ( ?, ?, ? )";
myAdapter.UpdateCommand.CommandText = "update MyTable set ColX = ?, ColY = ?, ColZ = ? where MyID = ?";
myAdapter.DeleteCommand.CommandText = "delete from MyTable where MyID = ?";

Now, each has to have their respective “Parameters”. The parameters have to be addded in the same sequence as their corresponding “?” place-holders.

// Although I’m putting in bogus values for preparing the parameters, its just for
// data type purposes. It does get changed through the data adapter when it applies the changes

OleDbParameter oParm = new OleDbParameter( "myID", -1 );
oParm.DbType = DbType.Int32;
oParm.SourceColumn = "myID";  // <- this is where it looks back to source table's column
oParm.ParameterName = "myID";  // just for consistency / readability reference

myAdapter.SelectCommand.Parameters.Add( oParm );

do similar for rest of parameters based on their types… char, int, double, whatever

Again, I have like a wrapper class that handles managment on a per-table basis… in brief

public myClassWrapper
{
    protected DataTable myTable;
    protected DataAdapter myAdapter;
    ... more ...

    protected void SaveChanges()
    {
    }
}

Its more complex than just this, but during the “SaveChanges”, The datatable and dataAdapter are in synch for their own purposes. Now, flushing the data. I check for the status of the table and then you can pass the entire table to the dataAdapter for update and it will cycle through all changed records and push respective changes. You’ll have to trap for whatever possible data errors though.

myAdapter.Update( this.MyTable );

As it finds each “changed” record, it pulls the values from the Column Source as identified by the parameter that is found in the table being passed to the adapter for processing.

Hopefully this has given you a huge jump on what you are running into.

—- COMMENT PER FEEDBACK —-

I would put your update within a try/catch, and step into the program to see what the exception is. The message adn/or inner exception of the error might give more info. However, try to simplify your UPDATE to only include a FEW fields with the WHERE “Key” element.

Additionally, and I oopsed, missed this from first part answer. You might have to identify the datatable’s “PrimaryKey” column. To do so, its a property of the DataTable that expects and array of columns that represent the primary key for the table. What I did was…

// set the primary key column of the table
DataColumn[] oCols = { myDataTbl.Columns["myID"] };
myDataTbl.PrimaryKey = oCols;

I would comment out your full update string and all its parameters for your UPDATE. Then, build it with just as simple as my sample of only setting 2-3 columns and the where clause

myAdapter.UpdateCommand.CommandText = "update MyTable set ColX = ?, ColY = ? where MyID=?";
Add Parameter object for "X"
Add Parameter object for "Y"
Add Parameter object for "MyID"

Pick fields like int or char so they have the least probability of problems for data type conversions, then, once that works, try adding all your “int” and “character” columns… then add any others. Also, which database are you going against. SOME databases don’t use “?” as placeholder in the command but use “named” parameters, some using

"actualColumn = @namedCol"
or even
"actualColumn = :namedCol"

Hope this gets you over the hump…