How skip rows and cells to create the dataset?

Hi guys.

We are using ExcelDataReader and ExcelDataReader.DataSet extension available on NuGet, to have the IExcelDataReader instance as DataSet. It works perfectly for the most of our scenarios and it show us such as a great tool.

There are some scenarios where we need to define the row and column (for sample: B3 or D4 or E10), where should start reading data and build the dataSet. For sample, I have worksheets where the data that interest me to extract start on F2 column (where F2 is the header line and above it the data).

Is there any way to specify what cell the DataSet should start reading using the configuration objects? Or how rows it should skip (for sample, 6 lines and start reading)?

Thank you.

[andersnm]

Hi!

Alas there is no way to exclude columns with AsDataSet currently.

But there are two options for skippings rows. The simplest is probably to use the FilterRow callback:

var dataset = reader.AsDataSet(new ExcelDataSetConfiguration() {
	ConfigureDataTable = _ => new ExcelDataTableConfiguration() {
		FilterRow = rowReader => rowReader.Depth > 6
	}
});

The above reads only rows when the Depth (row index) is greater than 6.

Can also skip rows before the header, with UseHeaderRow = true, and the ReadHeaderRow callback:

var dataset = reader.AsDataSet(new ExcelDataSetConfiguration() {
	ConfigureDataTable = _ => new ExcelDataTableConfiguration() {
                UseHeaderRow = true,
		ReadHeaderRow = rowReader => {
		            rowReader.Read();
		            rowReader.Read();
		            rowReader.Read();
		            rowReader.Read();
		            rowReader.Read();
		            rowReader.Read();
                }
	}
});

The above skips the 6 first rows, uses the next row as headers, then reads the rest as normal.

Which approach to use probably depends on the use case

[andersnm]

I think your first sample looks OK - it does what it needs to do. Alternatively, initialize the for-loop with the first column index to avoid the if-continue statement, f.ex for (int c = 1; c < table.Columns.Count; c++)...

The AsDataSet() options could use a new ExcelDataTableConfiguration.FilterColumn callback to configure which columns to read into the dataset. This would complement the FilterRow callback.

EDIT: Another request for the same https://github.com/ExcelDataReader/ExcelDataReader/issues/213#issuecomment-308731323