使用C ++将多个记录添加到ADO记录集的最快方法

Question

We have a data model that supports single table databases hosted on multiple servers. However we occasionally need to compile data from multiple tables for display on classic ASP pages. This is done in a COM object in C++. We gather the data from SQL Server using OleDb, with the OleDb logic writing the data straight into C++ structs. We then loop through the struct arrays adding the data to SafeArrays, then add these to the recordset. The logic to add the data to the recordset looks roughly like the following (initialisation and error handling logic omitted):

SAFEARRAY* fields; // Initialised to integer array 0 -> max fields
SAFEARRAY* values;
VARIANT* arraydata;

... //  initialisation logic

SafeArrayAccessData(values, reinterpret_cast<void**>(&arrayData));

for(unsigned int i = 0; i < numDataValues1; ++i)
{
    // Add data
    arraydata[0].intVal = data[i].someValue;
    ... // etc.

    for(unsigned int j = 0; j < numDataValues2; ++j)
    {
        arraydata[21].intVal = data2[j].someValue;
        ... // etc.

        _recordset->raw_AddNew(fields, values);
        _recordset->Update();
    }
}

However it is relatively slow. It can take a second to add a 1000 rows of data, which is just too long in our environment. If I comment out the calls to raw_AddNew and Update the logic flies. So the issue is with how I am adding the data to the recordset. Any suggestions or advice hugely appreciated. Big thanks in advance.

Answer 1

It has been a long time since I've coded ADO database connections, but if I remember right you need to change your locking to adLockBatchOptimistic and then do a batch update. It is much faster that way.

This might help you

If I remember right, it still was pretty slow.