Question

I have text files that are Tab delimited. I created a Schema.ini like so:

[MY_FILE.TAB]
Format=TabDelimited
ColNameHeader=False
Col1=id Short
Col2=data Text

This is the code I use to read it (C#):

using (var connection = new OleDbConnection(@"Provider=Microsoft.Jet.OLEDB.4.0;Data Source=D:FolderToData;Extended Properties= text;FMT=delimited "))
{
  using (var command = new OleDbCommand("SELECT * FROM MY_FILE.TAB", connection))
  {
    var table = new DataTable();
    using (var adapter = new OleDbDataAdapter(command)
    {
      adapter.Fill(table);
    }
  }
}

Everything works fine, except for one thing. The data in the text file contains Carriage Returns [CR]. Records itself are separated by Carriage Return Line Feeds [CR][LF]. Unfortunately, OleDB / MicrosoftJet (or whatever parses these files) treats both ([CR], [CRLF]) the same.

Example of MY_FILE.TAB (there should be a Tab between numbers and text):

1   One[CR][LF]
2   Two[CR][LF]
3   Th[CR]
ree[CR][LF]
4   Four[CR][LF]

Gives me 5 (malformed) Rows in the DataTable instead of 4.

What I need is:

1   "One"
2   "Two"
3   "Th
ree"
4   "Four2

But I get:

1    "One"
2    "Two"
3    "Th"
null null
4    "Four"

"ree" can t be converted to Int32 so first colum in fourth row is null.

How can I configure OleDB to treat [CR] different than [CR][LF]? Or any other Ideas?

Answer 1

I don t believe you can reconfigure OLEDB to do this directly.

An alternative approach would be to use a TextReader and TextWriter to process the file into a temporary file, scanning for and replacing CR alone into some special escape sequence. Then use OLEDB to read this replacement temporary file; finally, replace the special escape sequence back to a CR.

Answer 2

Wouldn t it be easy to read the file contents in a string, split it by Environment.NewLine or , which will get you an array for each line, which you can further split by tab?

友情链接