I see. Maybe this would be a better way to accomplish this. You could create an object called something like CellsInColumnInfo which would keep track of the information about the cells in the column as you were walking over data. Then you could loop over all created rows and their created cells, which would be the fastest way to loop over all the data. When you come across cells with data, add information about that cell to the CellsInColumnInfo for its column. Then after walking over all data, do something with your finalized CellsInColumnInfo instances, like this:

(Sorry about the code formatting)

Dictionary<int, CellsInColumnInfo> columnProperties = new Dictionary<int, CellsInColumnInfo>();
foreach (WorksheetRow row in workbook.Worksheets[0].Rows)
{
  foreach (WorksheetCell cell in row.Cells)
  {
   object cellValue = cell.Value;
   if (cellValue != null)
   {
    // scan all cell values to find the value properties e.g. biggest value in the column, is value numeric etc.
   CellsInColumnInfo columnProps;
    if (columnProperties.TryGetValue(cell.ColumnIndex, out columnProps) == false)
    {
     columnProps = new CellsInColumnInfo();
     columnProperties.Add(cell.ColumnIndex, columnProps);
    }
   // Store some information about the cells in the column
    columnProps.AddInformationAboutCell(cell, cellValue);
   }
  }
 }
foreach (KeyValuePair<int, CellsInColumnInfo> pair in columnProperties)
 {  
int columnIndex = pair.Key;
  CellsInColumnInfo columnProps = pair.Value;
 //column type decision code
 // on the basis of column properties determin appropriate sql data type   
// ...
}

thanks for the reply,
I am creating an application that according to the content of excel file creates SQL script for generating the tables on the SQL server.

result example:

CREATE TABLE <table_name>
(
ID int IDENTITY (1,1) Primary key,
[ProductID] bigint,
[ProductName] nvarchar(28),
[Column_03] datetime,
[UnitPrice] nvarchar(5),
[UnitsInStock] decimal(8,3),
[UnitsOnOrder] tinyint

...
)

The colum names and data types in sql script accurately reflect the contents of individual fields of Excel.

Application uses Infragistics dll.

It opens the excel file and have to scan through all values in the colum to determin correct sql server data type.

Example of code:

numberOfColumns = workbook.CurrentFormat == WorkbookFormat.Excel2007 ? 
Workbook.MaxExcel2007ColumnCount : Workbook.MaxExcelColumnCount;

for (int i = 0; i < numberOfColumns; i++) // here I need last used column in the file
{

WorksheetColumn column = workbook.Worksheets[0].Columns[i];

...

 foreach (var row in workbook.Worksheets[0].Rows)
 {

      if (row.GetCellValue(column.Index) != null)
          {
           columnValue = row.GetCellValue(column.Index).ToString().Trim();
          }

...

// scan all cell values to find the value properties e.g. biggest value in the column, is value numeric etc.

}

//column type decision code

// on the basis of column properties determin appropriate sql data type 

...
}

Mentioned code scans all columns (16384 columns) in excel file but I need to scan only used columns (e.g. tens). 

And my question is: How to determin the last used column in the excel worksheet?

Thank you