Question

I have no control over how the data is saved in this table. However, I have to query the table and combine the data for similar pn_id column as one row/record.

For instance current data structure is as follows,

Here we have same pn_id repeated with different question ids. This should have been really saved as one pn_id and then each question as a separate column, per my opinion. However, I have to retrieve the below data as one record like this this..

Any idea how this can be done?

Thanks

Answer 1

Here’s some pseudocode for the transform algorithm. Note that it requires scanning the entire data set twice; there are a few other opportunities to improve the efficiency, for example, if the input data can be sorted. Also, since it’s pseudocode, I haven’t added handling for null values.

var columnNames = new HashSet<string> { "pn_id" };

foreach (var record in data)
    columnNames.Add(record.question_id.ToString());

var table = new DataTable();
foreach (var name in columnNames)
    table.Columns.Add(new DataColumn(name, typeof(string)));

foreach (var record in data)
{
    var targetRecord = CreateNewOrGetExistingRecord(table, record.pn_id);
    targetRecord[record.question_id.ToString()] = record.char_value ?? record.date_value.ToString();
}

And here’s a sketch of the helper method:

DataRow CreateNewOrGetExistingRecord(DataTable table, object primaryKeyValue)
{
    var result = table.Find(primaryKeyValue);
    if (result != null)
        return result;

    //add code here to create a new row, add it to the table, and return it to the caller
}