使用 POI 读取 Word docx 中的书签替换书签内容（汉字或合并外部文档内容）

Posted 2022-10-11 catoop

通过操作书签可以实现 word 模板替换变量的功能场景，本文一下代码内容，直接可以复制使用正常编译运行。

添加 maven 依赖

    <dependencies>
        <dependency>
            <groupId>org.apache.poi</groupId>
            <artifactId>poi-ooxml</artifactId>
            <version>5.2.2</version>
        </dependency>
        <dependency>
            <groupId>org.apache.poi</groupId>
            <artifactId>poi-scratchpad</artifactId>
            <version>5.2.2</version>
        </dependency>
        <dependency>
            <groupId>com.deepoove</groupId>
            <artifactId>poi-tl</artifactId>
            <version>1.12.0</version>
        </dependency>
    </dependencies>

操作书签的代码

1、ShanhyXWPFDocumentMerge.java

package org.example;

import com.deepoove.poi.xwpf.NiceXWPFDocument;
import com.deepoove.poi.xwpf.XmlXWPFDocumentMerge;
import org.apache.xmlbeans.XmlOptions;
import org.apache.xmlbeans.impl.store.DomImpl;
import org.openxmlformats.schemas.wordprocessingml.x2006.main.CTBody;
import org.openxmlformats.schemas.wordprocessingml.x2006.main.CTP;
import org.w3c.dom.Node;

import java.lang.reflect.InvocationTargetException;
import java.lang.reflect.Method;
import java.util.ArrayList;
import java.util.Iterator;
import java.util.List;

/**
 * 文档合并
 *
 * @author shanhy
 */
public class ShanhyXWPFDocumentMerge extends XmlXWPFDocumentMerge 

    /**
     * 将一个文档合并到另外一个文档指定段落的相对位置
     *
     * @param source 当前文档
     * @param mergeIterator 被合入的文档
     * @param targetParagraphNode 目标段落Node
     * @param deleteTargetParagraph 是否删除目标段落自身
     * @return
     * @throws Exception
     */
    public ShanhyXWPFDocument mergeToParagraphBefore(ShanhyXWPFDocument source, Iterator<ShanhyXWPFDocument> mergeIterator, Node targetParagraphNode, boolean deleteTargetParagraph) throws Exception 
        CTBody body = source.getDocument().getBody();

        List<String> addParts = createMergeableStrings(source, mergeIterator);

        String[] startEnd = truncatedStartEndXmlFragment(body);

//        CTP mergedContainer = paragraph.getCTP();
        XmlOptions options = new XmlOptions();
        options.setUseSameLocale(((DomImpl.Dom) targetParagraphNode).locale());
        CTP mergedBody = CTP.Factory.parse(startEnd[0] + String.join("", addParts) + startEnd[1], options);
        // instead insert xml-fragment?

//        new XWPFParagraph(CTP.Factory.parse(mergedBody.getDomNode(), options), source);

        Node mergedContainerParentNode = targetParagraphNode.getParentNode();

        // 将引入文档的整体插入到目标段落之前
        mergedContainerParentNode.insertBefore(mergedBody.getDomNode(), targetParagraphNode);

        if(deleteTargetParagraph) 
            // 删除掉目标段落自身
            mergedContainerParentNode.removeChild(targetParagraphNode);
        

//        mergedContainer.getDomNode().appendChild(CTP.Factory.parse(mergedBody.getDomNode(), options).getDomNode());
//        mergedContainer.set(mergedBody);
//        String xmlText = truncatedOverlapWP(body);
//        body.set(CTBody.Factory.parse(xmlText));
//        return source.generate(true);
        return source;
    

    /**
     * 反射调用父类方法
     *
     * @param methodName
     * @param params
     * @throws NoSuchMethodException
     * @throws InvocationTargetException
     * @throws IllegalAccessException
     */
    private Object invokeSuperMethod(String methodName, Class<?>[] paramClasses, Object[] params) throws NoSuchMethodException, InvocationTargetException, IllegalAccessException 
        Method method = this.getClass().getSuperclass().getDeclaredMethod(methodName, paramClasses);
        method.setAccessible(true);
        return method.invoke(this, params);
    

    @SuppressWarnings("unchecked")
    private List<String> createMergeableStrings(ShanhyXWPFDocument source, Iterator<ShanhyXWPFDocument> mergeIterator) 
        try 
            Object obj = invokeSuperMethod("createMergeableStrings", new Class[]NiceXWPFDocument.class, Iterator.class, new Object[]source, mergeIterator);
            if (obj instanceof List<?>) 
                return (List<String>) obj;
            
         catch (NoSuchMethodException | InvocationTargetException | IllegalAccessException e) 
            throw new RuntimeException(e);
        
        return new ArrayList<>();
    

    private String[] truncatedStartEndXmlFragment(CTBody body) 
        try 
            Object obj = invokeSuperMethod("truncatedStartEndXmlFragment", new Class[]CTBody.class, new Object[]body);
            return (String[]) obj;
         catch (NoSuchMethodException | InvocationTargetException | IllegalAccessException e) 
            throw new RuntimeException(e);
        
    

    private String truncatedOverlapWP(CTBody body) 
        try 
            Object obj = invokeSuperMethod("truncatedOverlapWP", new Class[]CTBody.class, new Object[]body);
            return (String) obj;
         catch (NoSuchMethodException | InvocationTargetException | IllegalAccessException e) 
            throw new RuntimeException(e);
        
    

2、ShanhyXWPFDocument.java

package org.example;

import com.deepoove.poi.xwpf.NiceXWPFDocument;
import org.w3c.dom.Node;

import java.io.IOException;
import java.io.InputStream;
import java.util.Iterator;

/**
 * 处理 Docx 文档内容处理
 *
 * @author shanhy
 */
public class ShanhyXWPFDocument extends NiceXWPFDocument 

    public ShanhyXWPFDocument(InputStream in) throws IOException 
        super(in);
    

    /**
     * 将一个文档合并到另外一个文档指定段落的相对位置
     *
     * @param source 当前文档
     * @param mergeIterator 被合入的文档
     * @param targetParagraphNode 目标段落Node
     * @param deleteTargetParagraph 是否删除目标段落自身
     *
     * @throws Exception
     */
    public void mergeToParagraphBefore(ShanhyXWPFDocument source, Iterator<ShanhyXWPFDocument> mergeIterator, Node targetParagraphNode, boolean deleteTargetParagraph) throws Exception 
        new ShanhyXWPFDocumentMerge().mergeToParagraphBefore(this, mergeIterator, targetParagraphNode, deleteTargetParagraph);
    

3、ShanhyDocxBookmarkService.java

package org.example;

import org.apache.poi.xwpf.usermodel.XWPFDocument;
import org.apache.poi.xwpf.usermodel.XWPFParagraph;
import org.apache.poi.xwpf.usermodel.XWPFRun;
import org.apache.xmlbeans.XmlException;
import org.apache.xmlbeans.XmlOptions;
import org.apache.xmlbeans.impl.store.DomImpl;
import org.openxmlformats.schemas.wordprocessingml.x2006.main.CTP;
import org.w3c.dom.Node;
import org.w3c.dom.NodeList;

import java.io.IOException;
import java.io.OutputStream;
import java.util.ArrayList;
import java.util.Collections;
import java.util.List;
import java.util.Map;

/**
 * 书签替换处理类
 * 
 * @author shanhy
 */
public class ShanhyDocxBookmarkService 

    /**
     * 因为docx为xml格式的结构，一下为docx中定义的部分常量引用
     **/
    public static final String RUN_NODE_NAME = "w:r";
    public static final String TEXT_NODE_NAME = "w:t";
    public static final String BOOKMARK_START_TAG = "w:bookmarkStart";
    public static final String BOOKMARK_END_TAG = "w:bookmarkEnd";
    public static final String BOOKMARK_ID_ATTR_ID = "w:id";
    public static final String NODENAME_BODY = "w:body";
    public static final String NODENAME_PARAGRAPH = "w:p";
    public static final String BOOKMARK_ID_ATTR_NAME = "w:name";
    public static final String STYLE_NODE_NAME = "w:rPr";
    public static final String PARAGRAPH_PROPERTIES_NAME = "w:pPr";

    /**
     * 读取 docx 文件中的所有书签（注意不支持书签嵌套，书签嵌套书签的情况只识别最外层书签）
     *
     * @param docx
     */
    public List<Node> getBookmarksFromDocx(XWPFDocument docx) 
        Node bodyNode = docx.getDocument().getBody().getDomNode();

        // 递归读取 bookmarkStart 节点，返回bookmarkNode集合（<w:bookmarkStart w:id="1" w:name="书签名称"/>）
        List<Node> bookmarkNodeList = new ArrayList<>();
        getBookmarksFromNode(bodyNode, bookmarkNodeList);

        return bookmarkNodeList;
    

    /**
     * 递归解析所有Node节点，将bookmark记录到集合中
     *
     * @param node
     * @param bookmarkNodeList
     */
    public void getBookmarksFromNode(Node node, List<Node> bookmarkNodeList) 
        if (node.getNodeName().equals(BOOKMARK_START_TAG)) 
            bookmarkNodeList.add(node);
         else if (node.getNodeName().equals(BOOKMARK_END_TAG) || node.getNodeName().equals(PARAGRAPH_PROPERTIES_NAME)) 
            // Nothing
         else 
            NodeList childNodes = node.getChildNodes();
            for (int i = 0, j = childNodes.getLength(); i < j; i++) 
                Node childNode = childNodes.item(i);
                getBookmarksFromNode(childNode, bookmarkNodeList);
            
        
    

    /**
     * docx 文件中书签的替换
     *
     * @param docx
     * @param outputStream
     * @param dataMap
     * @throws IOException
     */
    public void replaceDocxBookmarks(ShanhyXWPFDocument docx, OutputStream outputStream, Map<String, Object> dataMap) throws Exception 
        // 获取所有书签
        List<Node> startBookmarkList = getBookmarksFromDocx(docx);

        // 替换书签内容
        for (Node startBookmarkNode : startBookmarkList) 
            String bookmarkName = startBookmarkNode.getAttributes().getNamedItem("w:name").getNodeValue();
            if (dataMap.containsKey(bookmarkName)) 
                Object data = dataMap.get(bookmarkName);
                if (data instanceof String) // 内容是文本
                    replaceDocxBookmarkFromString(getFirstParentParagraphByNode(startBookmarkNode, docx), startBookmarkNode, (String) data);
                 else if (data instanceof ShanhyXWPFDocument)  // 内容是外部 docx 文档
                    replaceDocxBookmarkFromDocx(docx, getFirstParentNodeByNode(startBookmarkNode, docx), startBookmarkNode, (ShanhyXWPFDocument) data);
                 else 
                    throw new RuntimeException("替换书签的内容源数据格式暂不支持");
                
            
        

        docx.write(outputStream);
        docx.close();
    

    /**
     * 将一个docx文档替换到docx的书签中
     * 1.要求书签尽量设置在文档的换行起点
     * 2.被合入的docx会作为换行起点开始合入，如果bookmark没有设置在一个换行的起点，程序会自动从该bookmark节点开始寻找，
     * 定位到父节点为body的对应p节点的下一个节点，然后将需要合入的文档所有内容追加到该p节点的下一个
     *
     * @param bookmarkParentParagraph
     * @param startBookmarkNode
     * @param content
     */
    public void replaceDocxBookmarkFromString(XWPFParagraph bookmarkParentParagraph, Node startBookmarkNode, String content) 
        Node nextNode = startBookmarkNode.getNextSibling();
        boolean contentReplaced = false;
        while (nextNode != null) 
            // 节点是结束符
            if (nextNode.getNodeName().equals(BOOKMARK_END_TAG)) 
                break;
            
            // 1.寻找startBookmark的下一个 w:r 节点，然后将该节点中的 w:t 节点的真正文字内容替换掉（这样可以保留原来的bookmark的文字样式）
            if (!contentReplaced && nextNode.getNodeName().equals("w:r")) 
                NodeList runChildNodes = nextNode.getChildNodes();
                for (int i = 0, j = runChildNodes.getLength(); i < j; i++) 
                    if (runChildNodes.item(i).getNodeName().equals("w:t")) // w:t 是真正的文本内容
                        runChildNodes.item(i).getFirstChild().setNodeValue(content);
                    
                
                contentReplaced = true;
             else 
                // 2.然后继续向下删除 endBookmark 之前的所有节点
                startBookmarkNode.getParentNode().removeChild(nextNode);
            
            nextNode = nextNode.getNextSibling();
        
        //1-end

        if (!contentReplaced) 
            XWPFRun run = bookmarkParentParagraph.createRun();
            run.setText(content);
            跪求POI根据模板生成word文档的例子。 注意啦，是用书签标记的方式，替换书签中的内容。

 将 RTF/DOC/DOCX 文件插入 Word 文件中的书签位置
 java poi 生成word表格怎么 让表格填充整个页面和合并单元格
 分别使用docx4j,jacob将文字与图片插入word中书签位置
 java用poi导出word文档，我要导出一个表格，表格的单元格中还要有一个表格，请问怎么实现
 java 用POI 解析word中的表格,POI只能识别word中创建的表格。 如果表格是从Excel中copy过来的, POI无法识