C#读取CSV文件异常求助：含表头/无表头文件重复打开问题

Hey Thomas, I see exactly what's causing those weird issues with your CSV loading—let's break it down and fix it step by step!

问题根源分析 #

Your current code only clears the rows of the DataTable with dt.Clear(), but it doesn't reset the columns between file openings. That's why you're getting inconsistent behavior:

• For CSV files with headers:
  • First open: Columns are created (using 1,2,3... as names), and the header row is skipped (since it's used to define column count). But when you open the file again, columns already exist, so the header line gets added as a data row—hence it shows up in the grid.
• For CSV files without headers:
  • First open: The first data line is used to create columns (so that row is lost, never added to the DataTable). Next time you open it, columns exist, so the first line is added as a row—so it shows up correctly.

修复方案 #

We need to fully reset the DataTable every time you open a new file, and add support for both header and non-header CSVs to handle your two file types properly.

Step 1: Fully reset the DataTable

Replace dt.Clear(); with a full reset to wipe both columns and rows:

dt.Reset();
// Or manually clear columns and rows: dt.Columns.Clear(); dt.Rows.Clear();

This ensures every file load starts with a blank slate, no leftover columns from previous files.

Step 2: Modified Code with Full Fix

Here's the updated code that handles both file types and fixes the reload issues:

private void btoDateiOeffnen_Click(object sender, EventArgs e)
{
    // Fully reset DataTable for each new file
    dt.Reset();

    if (oFDcsv.ShowDialog() == DialogResult.OK)
    {
        // Let user specify if the CSV has headers
        bool hasHeaders = MessageBox.Show("Does this CSV have column headers?", "CSV Option", MessageBoxButtons.YesNo) == DialogResult.Yes;

        using (StreamReader sr = new StreamReader(oFDcsv.FileName, Encoding.UTF8))
        {
            string firstLine = sr.ReadLine()?.Replace(",", ".");
            if (string.IsNullOrEmpty(firstLine))
            {
                MessageBox.Show("Empty CSV file!");
                return;
            }
            string[] firstLineParts = firstLine.Split(new string[] { ";" }, StringSplitOptions.None);

            // Create columns based on user choice
            for (int i = 0; i < firstLineParts.Length; i++)
            {
                string columnName = hasHeaders ? firstLineParts[i] : Convert.ToString(i + 1);
                dt.Columns.Add(columnName);
            }

            // Add first line as data if there are no headers
            if (!hasHeaders)
            {
                dt.Rows.Add(firstLineParts);
            }

            // Read and add remaining lines
            while (!sr.EndOfStream)
            {
                string[] currentLine = sr.ReadLine().Replace(",", ".").Split(new string[] { ";" }, StringSplitOptions.None);
                dt.Rows.Add(currentLine);
            }
        }

        // Update DataGridView
        dgvDateiAnsicht.DataSource = dt;
        
        // Hide third column only if it exists
        if (dgvDateiAnsicht.Columns.Count >= 3)
        {
            dgvDateiAnsicht.Columns[2].Visible = false;
        }
        
        // Disable sorting for all columns
        foreach (DataGridViewColumn column in dgvDateiAnsicht.Columns)
        {
            column.SortMode = DataGridViewColumnSortMode.NotSortable;
        }
    }
}

关键改进点 #

• Full DataTable Reset: Ensures no leftover columns from previous file loads.
• Header Toggle: Lets users specify if their CSV has headers, so:
  • Header CSVs: First line becomes column names, remaining lines are data.
  • Non-header CSVs: First line is added as data, columns get numeric names.
• Error Prevention: Added checks for empty files and missing columns before hiding the third column.
• Cleaner Resource Handling: Removed redundant sr.Close()—the using block automatically disposes the StreamReader.

This should fix both of your issues—no more missing first rows or extra header rows when reloading files!

内容的提问来源于stack exchange，提问作者Thomas