From b44bfff58de3adc572fd4643ecb3ec9b5586bea3 Mon Sep 17 00:00:00 2001
From: 许鹏程 <1821349743@qq.com>
Date: 星期二, 28 五月 2024 16:09:03 +0800
Subject: [PATCH] commit
---
src/main/java/com/product/lucene/util/FileUtils.java | 253 ++++++++++++++++++++++++++++++++-----------------
1 files changed, 164 insertions(+), 89 deletions(-)
diff --git a/src/main/java/com/product/lucene/util/FileUtils.java b/src/main/java/com/product/lucene/util/FileUtils.java
index be965f3..c7fd7ca 100644
--- a/src/main/java/com/product/lucene/util/FileUtils.java
+++ b/src/main/java/com/product/lucene/util/FileUtils.java
@@ -8,12 +8,14 @@
import java.io.InputStream;
import java.util.List;
+import com.product.common.lang.StringUtils;
import org.apache.poi.hssf.usermodel.HSSFCell;
import org.apache.poi.hssf.usermodel.HSSFRow;
import org.apache.poi.hssf.usermodel.HSSFSheet;
import org.apache.poi.hssf.usermodel.HSSFWorkbook;
import org.apache.poi.hwpf.HWPFDocument;
import org.apache.poi.hwpf.usermodel.Range;
+import org.apache.poi.ss.usermodel.CellType;
import org.apache.poi.xssf.usermodel.XSSFCell;
import org.apache.poi.xssf.usermodel.XSSFRow;
import org.apache.poi.xssf.usermodel.XSSFSheet;
@@ -28,83 +30,139 @@
public class FileUtils {
/**
- * 鏂囦欢杞瑂tring
+ * 鏂囦欢杞瑂tring
+ *
* @param file
* @return
*/
public static String FileToString(File file) {
- String file_content="";
- if ((file != null) && (file.isFile())) {
- if (file.getName().toLowerCase().endsWith(".txt")) {
- file_content = txtToString(file);
- } else if ((file.getName().toLowerCase().endsWith(".doc"))){
- file_content = docToString(file);
- } else if ((file.getName().toLowerCase().endsWith(".docx"))) {
- file_content = docxToString(file);
- } else if (file.getName().toLowerCase().endsWith(".xls")) {
- file_content = readXls(file);
- } else if (file.getName().toLowerCase().endsWith(".xlsx")) {
- file_content = readXlsx(file);
- }
+ String file_content = "";
+ String fileType = "unknown";
+ //鑾峰彇鏂囦欢鍚庣紑 淇濈暀.鍙�
+ String fileSuffix = file.getName().substring(file.getName().lastIndexOf("."));
+ if (StringUtils.equalsAny(fileSuffix, ".doc", ".docx", ".xls", ".xlsx")) {
+ //璇诲彇鏂囦欢
+ fileType = checkDocType(file);
}
- return file_content;
- }
-
+ try {
+ if ((file != null) && (file.isFile())) {
+
+
+ if ("doc".equals(fileType)) {
+ file_content = docToString(file);
+ } else if ("docx".equals(fileType)) {
+ file_content = docxToString(file);
+ } else if ("xls".equals(fileType)) {
+ file_content = readXls(file);
+ } else if ("xlsx".equals(fileType)) {
+ file_content = readXlsx(file);
+ } else if (file.getName().toLowerCase().endsWith(".txt")) {
+ file_content = txtToString(file);
+ }
+ }
+ } catch (Exception e) {
+ e.printStackTrace();
+ }
+ return file_content;
+ }
+
+
+ public static String checkDocType(File file) {
+ try (FileInputStream fis = new FileInputStream(file)) {
+ byte[] bytes = new byte[8];
+ fis.read(bytes, 0, 8);
+
+ String hex = bytesToHex(bytes);
+
+ if (hex.contains("504B0304") && (file.getName().endsWith(".doc") || file.getName().endsWith(".docx"))) {
+ return "docx";
+ } else if (hex.contains("D0CF11E0") && (file.getName().endsWith(".doc") || file.getName().endsWith(".docx"))) {
+ //鍥犱负doc鏂囦欢鐨勫ご閮ㄤ篃鏄疍0CF11E0锛屾墍浠ラ渶瑕佸垽鏂枃浠跺悗缂�
+ return "doc";
+ }
+ //澧炲姞xls 鍜� xlsx鐨勫垽鏂�
+ else if (hex.contains("504B0304") && (file.getName().endsWith(".xls") || file.getName().endsWith(".xlsx"))) {
+ return "xlsx";
+ } else if (hex.contains("D0CF11E0") && (file.getName().endsWith(".xls") || file.getName().endsWith(".xlsx"))) {
+ //鍥犱负xls鏂囦欢鐨勫ご閮ㄤ篃鏄疍0CF11E0锛屾墍浠ラ渶瑕佸垽鏂枃浠跺悗缂�
+ return "xls";
+ } else {
+ return "unknown";
+ }
+ } catch (
+ IOException e) {
+ e.printStackTrace();
+ return "unknown";
+ }
+ }
+
+ private static String bytesToHex(byte[] bytes) {
+ StringBuilder hex = new StringBuilder();
+ for (byte b : bytes) {
+ hex.append(String.format("%02X", b));
+ }
+ return hex.toString();
+ }
+
/**
- * txt鏂囦欢璇诲彇
+ * txt鏂囦欢璇诲彇
+ *
* @param file
* @return
*/
- public static String txtToString(File file) {
+ public static String txtToString(File file) throws IOException {
String result = "";
- try (BufferedReader br = new BufferedReader(new FileReader(file))){// 鏋勯�犱竴涓狟ufferedReader绫绘潵璇诲彇鏂囦欢
+ try (BufferedReader br = new BufferedReader(new FileReader(file))) {// 鏋勯�犱竴涓狟ufferedReader绫绘潵璇诲彇鏂囦欢
String s = null;
// 浣跨敤readLine鏂规硶锛屼竴娆¤涓�琛�
while ((s = br.readLine()) != null) {
result = result + "\n" + s;
}
} catch (Exception e) {
- e.printStackTrace();
+ throw e;
}
return result;
}
-
+
/**
- * doc杞枃鏈�
+ * doc杞枃鏈�
+ *
* @param file
* @return
*/
- public static String docToString(File file) {
+ public static String docToString(File file) throws IOException {
String result = "";
try (
- FileInputStream fileInputStream = new FileInputStream(file);
- HWPFDocument doc = new HWPFDocument(fileInputStream);
- ){
+ FileInputStream fileInputStream = new FileInputStream(file);
+ HWPFDocument doc = new HWPFDocument(fileInputStream);
+ ) {
Range rang = doc.getRange();
result += rang.text();
} catch (Exception e) {
- e.printStackTrace();
+ throw e;
}
return result;
}
-
+
/**
- * docx杞枃鏈�
+ * docx杞枃鏈�
+ *
* @param file
* @return
* @throws IOException
*/
- public static String docxToString(File file) {
- StringBuffer s=new StringBuffer();
+ public static String docxToString(File file) throws IOException {
+
+ StringBuffer s = new StringBuffer();
try (
- InputStream inputStream = new FileInputStream(file);
- XWPFDocument doc = new XWPFDocument(inputStream);
- ){
+ InputStream inputStream = new FileInputStream(file);
+ XWPFDocument doc = new XWPFDocument(inputStream);
+ ) {
List<XWPFParagraph> paras = doc.getParagraphs();
for (XWPFParagraph para : paras) {
// 褰撳墠娈佃惤鐨勫睘鎬�
// CTPPr pr = para.getCTP().getPPr();
- s.append( para.getText());
+ s.append(para.getText());
}
// 鑾峰彇鏂囨。涓墍鏈夌殑琛ㄦ牸
List<XWPFTable> tables = doc.getTables();
@@ -119,29 +177,29 @@
// 鑾峰彇琛屽搴旂殑鍗曞厓鏍�
cells = row.getTableCells();
for (XWPFTableCell cell : cells) {
- s.append( cell.getText());
+ s.append(cell.getText());
}
}
}
} catch (Exception e) {
- e.printStackTrace();
+ throw e;
}
return s.toString();
}
-
+
/**
* `xls杞瓧绗�
- * @param f
+ *
+ * @param file
* @return
* @throws IOException
*/
- public static String readXls(File file) {
+ public static String readXls(File file) throws IOException {
StringBuffer info = new StringBuffer();
- try(
- InputStream inputStream = new FileInputStream(file);
- HSSFWorkbook hssfWorkbook = new HSSFWorkbook(inputStream);
- )
- {
+ try (
+ InputStream inputStream = new FileInputStream(file);
+ HSSFWorkbook hssfWorkbook = new HSSFWorkbook(inputStream);
+ ) {
// Read the Sheet
for (int numSheet = 0; numSheet < hssfWorkbook.getNumberOfSheets(); numSheet++) {
HSSFSheet hssfSheet = hssfWorkbook.getSheetAt(numSheet);
@@ -161,88 +219,105 @@
}
}
} catch (Exception e) {
- e.printStackTrace();
+ throw e;
}
return info.toString();
}
-
+
/**
- * xlsx杞瓧绗�
+ * xlsx杞瓧绗�
+ *
* @param file
* @return
* @throws IOException
*/
- public static String readXlsx(File file) {
+ public static String readXlsx(File file) throws IOException {
StringBuffer info = new StringBuffer();
- try(
+ try (
InputStream is = new FileInputStream(file);
XSSFWorkbook xssfWorkbook = new XSSFWorkbook(is);
- ){
-
- for (int numSheet = 0; numSheet < xssfWorkbook.getNumberOfSheets(); numSheet++) {
- XSSFSheet xssfSheet = xssfWorkbook.getSheetAt(numSheet);
- if (xssfSheet == null) {
- continue;
- }
- // Read the Row
- for (int rowNum = 0; rowNum <= xssfSheet.getLastRowNum(); rowNum++) {
- XSSFRow xssfRow = xssfSheet.getRow(rowNum);
- if (xssfRow != null) {
- int tdLength = xssfRow.getLastCellNum();
- for (int j = 0; j <= tdLength; j++) {
- XSSFCell no = xssfRow.getCell(j);
- info.append(getValue(no)).append(" ");
+ ) {
+
+ for (int numSheet = 0; numSheet < xssfWorkbook.getNumberOfSheets(); numSheet++) {
+ XSSFSheet xssfSheet = xssfWorkbook.getSheetAt(numSheet);
+ if (xssfSheet == null) {
+ continue;
+ }
+ // Read the Row
+ for (int rowNum = 0; rowNum <= xssfSheet.getLastRowNum(); rowNum++) {
+ XSSFRow xssfRow = xssfSheet.getRow(rowNum);
+ if (xssfRow != null) {
+ int tdLength = xssfRow.getLastCellNum();
+ for (int j = 0; j <= tdLength; j++) {
+ XSSFCell no = xssfRow.getCell(j);
+ info.append(getValue(no)).append(" ");
+ }
}
}
}
- }
} catch (Exception e) {
- e.printStackTrace();
+ throw e;
}
return info.toString();
}
-
+
/**
- * 鑾峰彇xls鍗曞厓鏍煎��
+ * 鑾峰彇xls鍗曞厓鏍煎��
+ *
* @param hssfCell
* @return
*/
private static String getValue(HSSFCell hssfCell) {
- if (hssfCell == null)
+ if (hssfCell == null) {
return "";
- if (hssfCell.getCellType() == hssfCell.CELL_TYPE_BOOLEAN) {
- return String.valueOf(hssfCell.getBooleanCellValue());
- } else if (hssfCell.getCellType() == hssfCell.CELL_TYPE_NUMERIC) {
- return String.valueOf(hssfCell.getNumericCellValue());
- } else {
- hssfCell.setCellType(hssfCell.CELL_TYPE_STRING);
- return String.valueOf(hssfCell.getStringCellValue());
+ }
+ CellType cellType = hssfCell.getCellType();
+ //濡傛灉涓嶆槸瀛楃绫诲瀷鍒欒繑鍥炵┖
+ if (!CellType.BOOLEAN.equals(cellType) && !CellType.NUMERIC.equals(cellType) && !CellType.STRING.equals(cellType)) {
+ return "";
+ }
+ try {
+ if (CellType.BOOLEAN.equals(cellType)) {
+ return String.valueOf(hssfCell.getBooleanCellValue());
+ } else if (CellType.NUMERIC.equals(cellType)) {
+ return String.valueOf(hssfCell.getNumericCellValue());
+ } else {
+ return String.valueOf(hssfCell.getStringCellValue());
+ }
+ } catch (Exception e) {
+ throw e;
}
}
-
+
/**
- * 鑾峰彇xlxs鍗曞厓鏍煎��
+ * 鑾峰彇xlxs鍗曞厓鏍煎��
+ *
* @param xssfRow
* @return
*/
private static String getValue(XSSFCell xssfRow) {
if (xssfRow == null)
return "";
- if (xssfRow.getCellType() == xssfRow.CELL_TYPE_BOOLEAN) {
+ CellType cellType = xssfRow.getCellType();
+ //濡傛灉涓嶆槸瀛楃绫诲瀷鍒欒繑鍥炵┖
+ if (!CellType.BOOLEAN.equals(cellType) && !CellType.NUMERIC.equals(cellType) && !CellType.STRING.equals(cellType)) {
+ return "";
+ }
+ if (CellType.BOOLEAN.equals(cellType)) {
return String.valueOf(xssfRow.getBooleanCellValue());
- } else if (xssfRow.getCellType() == xssfRow.CELL_TYPE_NUMERIC) {
+ } else if (CellType.NUMERIC.equals(cellType)) {
return String.valueOf(xssfRow.getNumericCellValue());
} else {
return String.valueOf(xssfRow.getStringCellValue());
}
}
-
- public static void main(String[] args) {
- File f1=new File("D://ceshi/a.txt");
- File f2=new File("D://ceshi/b.doc");
- File f3=new File("D://ceshi/c.docx");
- File f4=new File("D://ceshi/d.xls");
- File f5=new File("D://ceshi/e.xlsx");
+
+ public static void main(String[] args) throws Exception {
+ File f1 = new File("D://ceshi/a.txt");
+ File f2 = new File("D://ceshi/b.doc");
+ File f3 = new File("D://ceshi/c.docx");
+ File f4 = new File("D://ceshi/d.xls");
+ File f5 = new File("D://ceshi/e.xlsx");
System.out.println(txtToString(f1));
System.out.println("+++++++++++++++++++++++++++++++++++++++++++");
System.out.println(docToString(f2));
--
Gitblit v1.9.2