使用Docx4j将HTML转换为Docx时跨页表格断裂及文档全局属性、单元格间距问题求助

Let's break down solutions for each of your problems using Docx4j 8.3.2 and Java 8:

1. Fix Table Cross-Page Breakage #

Tables splitting across pages can be resolved either by adding CSS to your HTML or post-processing the imported DOCX content. Here's the post-processing approach (more reliable if CSS isn't fully supported):

WordprocessingMLPackage wordMLPackage = WordprocessingMLPackage.createPackage();
XHTMLImporterImpl XHTMLImporter = new XHTMLImporterImpl(wordMLPackage);
List<Object> content = XHTMLImporter.convert(file, null);

// Post-process tables to prevent splitting
for (Object obj : content) {
    if (obj instanceof Tbl) {
        Tbl table = (Tbl) obj;
        // Prevent entire table from splitting
        TblPr tblPr = table.getTblPr() != null ? table.getTblPr() : new TblPr();
        CantSplit cantSplit = new CantSplit();
        cantSplit.setVal(Boolean.TRUE);
        tblPr.getCantSplitOrTblWOrTblCaptionOrTblDescription().add(cantSplit);
        table.setTblPr(tblPr);

        // Optional: Keep individual rows together
        for (Object trObj : table.getContent()) {
            if (trObj instanceof Tr) {
                Tr row = (Tr) trObj;
                TrPr trPr = row.getTrPr() != null ? row.getTrPr() : new TrPr();
                
                KeepLines keepLines = new KeepLines();
                keepLines.setVal(Boolean.TRUE);
                trPr.getKeepLinesOrKeepNextOrNoWrap().add(keepLines);
                
                KeepNext keepNext = new KeepNext();
                keepNext.setVal(Boolean.TRUE);
                trPr.getKeepLinesOrKeepNextOrNoWrap().add(keepNext);
                
                row.setTrPr(trPr);
            }
        }
    }
}

wordMLPackage.getMainDocumentPart().getContent().addAll(content);
File fileDos = new File(urlWord);
wordMLPackage.save(fileDos); // Don't forget to save!

2. Set Global Paragraph Properties #

To modify global styles like paragraph spacing, update the default "Normal" style in your document:

// After creating the WordprocessingMLPackage
Styles styles = wordMLPackage.getMainDocumentPart().getStyleDefinitionsPart().getJaxbElement();
Style normalStyle = null;

// Find the Normal style
for (Style style : styles.getStyle()) {
    if ("Normal".equals(style.getStyleId())) {
        normalStyle = style;
        break;
    }
}

if (normalStyle != null) {
    PPr pPr = normalStyle.getPPr() != null ? normalStyle.getPPr() : new PPr();
    Spacing spacing = new Spacing();
    
    // Set values in twips (1 twip = 1/20 pt)
    spacing.setBefore(BigInteger.valueOf(120)); // 6pt before paragraph
    spacing.setAfter(BigInteger.valueOf(120));  // 6pt after paragraph
    spacing.setLine(BigInteger.valueOf(240));   // 1.5x line spacing (160 = 1x)
    spacing.setLineRule(STLineSpacingRule.AUTO);
    
    pPr.setSpacing(spacing);
    normalStyle.setPPr(pPr);
}

3. Fix Inconsistent Table Cell Spacing #

Inconsistencies often come from default Word cell margins or inconsistent HTML styles. Force consistent margins for all tables with this post-processing step:

// Add this loop after converting HTML to content (same as the table break fix)
for (Object obj : content) {
    if (obj instanceof Tbl) {
        Tbl table = (Tbl) obj;
        TblPr tblPr = table.getTblPr() != null ? table.getTblPr() : new TblPr();
        
        TblCellMar cellMar = new TblCellMar();
        // Set 2.5pt margin on all sides (50 twips = 2.5pt)
        TblWidth marginWidth = new TblWidth();
        marginWidth.setW(BigInteger.valueOf(50));
        marginWidth.setType(STTblWidth.DXA);
        
        cellMar.setTop(marginWidth);
        cellMar.setBottom(marginWidth);
        cellMar.setLeft(marginWidth);
        cellMar.setRight(marginWidth);
        
        tblPr.getTblCellMarOrTblBordersOrTblLook().add(cellMar);
        table.setTblPr(tblPr);
    }
}

Bonus Check: #

If the issue persists, verify your HTML has consistent padding/margin styles on <td> elements (e.g., <td style="padding: 2.5pt;">). The post-processing above will override most inconsistent HTML styles, though.

内容的提问来源于stack exchange，提问作者MartinGam