public static DataColumnReport CreateColumnReport(string fileName, bool containsHeader, char delimiter, int firstK)
{
if (!File.Exists(fileName))
{
throw new ArgumentException(String.Format("File {0} does not exist", fileName));
}
IDictionary <string, int> columnToOrder = null;
IDictionary <int, Histogram <ColumnNativeType> > histograms;
var cnt = 0;
using (var stream = new FileStream(fileName, FileMode.Open, FileAccess.Read))
{
using (var reader = new StreamReader(stream))
{
var colCount = -1;
var line = String.Empty;
if (containsHeader)
{
line = reader.ReadLine();
if (!String.IsNullOrEmpty(line))
{
columnToOrder = GetColumnOrder(line, delimiter);
colCount = columnToOrder.Count;
}
}
else
{
line = reader.ReadLine();
if (!String.IsNullOrEmpty(line))
{
colCount = line.Split(delimiter).Length;
if (stream.CanSeek)
{
stream.Seek(0, SeekOrigin.Begin);
}
else
{
throw new InvalidOperationException(String.Format("File {0} cant not be seeked", fileName));
}
}
}
if (colCount <= 0)
{
throw new ArgumentException(String.Format("File {0} is empty", fileName));
}
histograms = InitializeHistograms(colCount);
cnt = 0;
while ((line = reader.ReadLine()) != null)
{
var parsed = line.Split(delimiter);
if (parsed == null || parsed.Length != colCount)
{
throw new InvalidOperationException(String.Format("There is a non-conformant record in line {0} of file {1} ", cnt, fileName));
}
for (int i = 0; i < colCount; i++)
{
var nativeType = DataColumn.GetNativeType(parsed[i]);
var hist = histograms[i];
if (hist.ContainsKey(nativeType))
{
++hist[nativeType];
}
else
{
hist.Add(nativeType, 1);
}
}
++cnt;
if (cnt >= firstK)
{
break;
}
}
}
}
var orderToColumn = new Dictionary <int, string>();
foreach (var item in columnToOrder)
{
orderToColumn.Add(item.Value, item.Key);
}
var collection = new DataColumnCollection();
var listInvalid = new List <string>();
foreach (var item in histograms)
{
var order = item.Key;
var colName = orderToColumn[order];
var hist = item.Value;
var kv = InferNativeType(hist);
if (kv.Key)
{
var nativeType = kv.Value;
var measurementType = DataColumn.GetDefaultMeasurementType(nativeType);
var dc = new DataColumn(colName, order)
{
NativeType = nativeType, MeasurementType = measurementType
};
collection.Add(colName, dc);
}
else
{
listInvalid.Add(colName);
}
}
return(new DataColumnReport {
ColumnCollection = collection, InvalidColumns = listInvalid, ScannedNumOfLines = cnt
});
}