Legacy .doc files

Word 97–2003 binary documents load through the same calls as .docx. The format is detected from the file's content, not its extension. Convert them to PDF, or open and save them as .docx to modernise them. There is no .doc writer.

Read, convert, modernise

using DocWright;
using DocWright.Core.Diagnostics;
using DocWright.Dom;

var converter = new DocWrightConverter();
var diagnostics = new CollectingConversionDiagnostics();

// A .doc is detected from its content, like everything else.
using FileStream input = File.OpenRead("legacy-tables.doc");
using WordDocument document = converter.Load(input, options: new ConvertOptions { Diagnostics = diagnostics });

Console.WriteLine($"format:   {document.SourceFormatName}");
Console.WriteLine($"sections: {document.Sections.Count}, styles: {document.Styles.Count}");

// Modernise: save as .docx. There is no .doc writer.
using (FileStream docx = File.Create("legacy-tables.docx"))
{
    converter.Save(document, docx, ".docx");
}

using (FileStream pdf = File.Create("legacy-tables.pdf"))
{
    converter.Convert(document, pdf);
}

foreach (ConversionDiagnostic d in diagnostics.Snapshot())
{
    Console.WriteLine($"{d.Code}: {d.Message[..Math.Min(90, d.Message.Length)]}…");
}

Output

format:   doc
sections: 1, styles: 6
DXP2002: MS-DOC table paragraphs were detected but full table reconstruction is not implemented; ce…
DXP2002: MS-DOC formatting: 242 sprm(s) across 47 distinct opcode(s) are not mapped and were skippe…
DXP2002: MS-DOC formatting: 2 property group(s) could not be read to their end and were applied as …

legacy-tables.doc → PDF

The table's rows, cells and column positions are read. Its borders and shading are not yet: compare the same document's .docx version.

What is read

Content	Read	Notes
Text, fields, hyperlinks, tabs	✅
Character formatting	✅	Font, size, bold, italic, strike, caps, small caps, hidden
Paragraph formatting	✅	Alignment, indents, spacing, keep-with-next, page break before
Styles	✅	With inheritance
Sections	✅	Page size, orientation, margins, columns
Tables	✅	Rows, cells and column grid. Nested tables are flattened into the outer cell.
Table borders and shading	❌	Not read yet
Pictures	✅	JPEG, PNG and DIB, with scaling and cropping
List numbering	❌	List text survives, the numbers don't
Headers and footers	❌	Reported, not read
Encryption	✅	Office 97 RC4 and RC4 CryptoAPI. Pass ConvertOptions.Password.

The diagnostics are normal

A .doc stores formatting as compact binary instructions ("sprms"). Those DocWright doesn't map are skipped, counted, and reported once per document as DXP2002, with their opcodes, as in the output above. Nearly every real document sets properties such as language that don't affect the page, so a non-empty list is expected. Treat it as a lead when investigating a specific difference, not as an error.