基于Html Agility Pack提取无标识表格数据的技术咨询

Got it, let's solve this problem using Html Agility Pack—since your tables don't have fixed IDs or XPaths, we'll anchor everything to that reliable CONSTANT_TEXT marker. Here's how to hit both your requirements:

1. Get the CHANGING_VALUE from the same row as CONSTANT_TEXT #

Instead of targeting the anchor tag first, start by finding the <td> that contains your fixed CONSTANT_TEXT. Once you have that node, you can easily grab the second <td> in its parent row, no matter where the table sits in the HTML structure.

// Locate the <td> holding the fixed CONSTANT_TEXT
var constantTd = document.DocumentNode.SelectSingleNode("//td[contains(text(), 'CONSTANT_TEXT')]");

if (constantTd != null)
{
    // Navigate up to the parent <tr> (the row containing our target td)
    var parentRow = constantTd.ParentNode;
    
    if (parentRow != null && parentRow.Name.Equals("tr", StringComparison.OrdinalIgnoreCase))
    {
        // Grab the second <td> in this row to get CHANGING_VALUE
        var valueTd = parentRow.SelectSingleNode("./td[2]");
        if (valueTd != null)
        {
            string changingValue = valueTd.InnerText.Trim();
            // Use your extracted value here
            Console.WriteLine($"Extracted value: {changingValue}");
        }
    }
}

2. Get the parent table containing CONSTANT_TEXT #

To find the closest (lowest-level) parent table, you can either traverse up the node tree manually or use an XPath ancestor axis for a cleaner approach:

Option 1: Manual traversal #

if (constantTd != null)
{
    HtmlNode parentTable = null;
    var currentNode = constantTd.ParentNode;
    
    // Move up through parent nodes until we hit a <table>
    while (currentNode != null && !currentNode.Name.Equals("table", StringComparison.OrdinalIgnoreCase))
    {
        currentNode = currentNode.ParentNode;
    }
    
    if (currentNode != null)
    {
        parentTable = currentNode;
        // Access the table's content, e.g., get its full HTML
        string tableHtml = parentTable.OuterHtml;
        Console.WriteLine($"Parent table HTML: {tableHtml}");
    }
}

Option 2: XPath ancestor axis (more concise) #

This XPath query directly finds the nearest parent <table> to your target <td>:

if (constantTd != null)
{
    var parentTable = constantTd.SelectSingleNode("./ancestor::table[1]");
    if (parentTable != null)
    {
        // Work with the parent table here
        Console.WriteLine($"Found parent table with {parentTable.SelectNodes(".//tr").Count} rows");
    }
}

Why this works better than your original approach #

Your initial code targeted the anchor tag with CONSTANT_ID, but since your core marker is CONSTANT_TEXT, starting there ensures you're always aligned with the right row and table—even if the anchor tag's ID or position shifts slightly across files.

内容的提问来源于stack exchange，提问作者michal