优化读写性能

README.md

@@ -21,7 +21,7 @@ EasyExcel
 # JAVA解析Excel工具EasyExcel
 Java解析、生成Excel比较有名的框架有Apache poi、jxl。但他们都存在一个严重的问题就是非常的耗内存，poi有一套SAX模式的API可以一定程度的解决一些内存溢出的问题，但POI还是有一些缺陷，比如07版Excel解压缩以及解压后存储都是在内存中完成的，内存消耗依然很大。easyexcel重写了poi对07版Excel的解析，一个3M的excel用POI sax解析依然需要100M左右内存，改用easyexcel可以降低到几M，并且再大的excel也不会出现内存溢出；03版依赖POI的sax模式，在上层做了模型转换的封装，让使用者更加简单方便
 
-## 64M内存1分钟内读取75M(46W行25列)的Excel
+## 64M内存20秒内读取75M(46W行25列)的Excel（3.0.2+版本）
 当然还有极速模式能更快，但是内存占用会在100M多一点
 ![img](img/readme/large.png)
 
@@ -41,7 +41,7 @@ Java解析、生成Excel比较有名的框架有Apache poi、jxl。但他们都
         <dependency>
             <groupId>com.alibaba</groupId>
             <artifactId>easyexcel</artifactId>
-            <version>3.0.1</version>
+            <version>3.0.2</version>
         </dependency>
 ```
 

pom.xml

@@ -4,7 +4,7 @@
     <modelVersion>4.0.0</modelVersion>
     <groupId>com.alibaba</groupId>
     <artifactId>easyexcel</artifactId>
-    <version>3.0.1</version>
+    <version>3.0.2</version>
     <packaging>jar</packaging>
     <name>easyexcel</name>
 

src/main/java/com/alibaba/excel/metadata/format/DataFormatter.java

@@ -111,6 +111,8 @@ public class DataFormatter {
      */
     private static final Pattern alternateGrouping = Pattern.compile("([#0]([^.#0])[#0]{3})");
 
+    private static final Pattern E_NOTATION_PATTERN = Pattern.compile("E(\\d)");
+
     /**
      * Cells formatted with a date or time format and which contain invalid date or time values show 255 pound signs
      * ("#").
@@ -644,8 +646,7 @@ public class DataFormatter {
      */
     private String getFormattedNumberString(BigDecimal data, Short dataFormat, String dataFormatString) {
         Format numberFormat = getFormat(data.doubleValue(), dataFormat, dataFormatString);
-        String formatted = numberFormat.format(data);
-        return formatted.replaceFirst("E(\\d)", "E+$1"); // to match Excel's E-notation
+        return E_NOTATION_PATTERN.matcher(numberFormat.format(data)).replaceFirst("E+$1");
     }
 
     /**

src/test/java/com/alibaba/easyexcel/test/core/large/LargeDataTest.java

@@ -32,6 +32,7 @@ public class LargeDataTest {
     private static File template07;
     private static File fileCsv;
     private static File fileWrite07;
+    private static File fileWriteTemp07;
     private static File fileWritePoi07;
 
     private int i = 0;
@@ -40,14 +41,14 @@ public class LargeDataTest {
     public static void init() {
         fileFill07 = TestFileUtil.createNewFile("largefill07.xlsx");
         fileWrite07 = TestFileUtil.createNewFile("large" + File.separator + "fileWrite07.xlsx");
+        fileWriteTemp07 = TestFileUtil.createNewFile("large" + File.separator + "fileWriteTemp07.xlsx");
         fileWritePoi07 = TestFileUtil.createNewFile("large" + File.separator + "fileWritePoi07.xlsx");
         template07 = TestFileUtil.readFile("large" + File.separator + "fill.xlsx");
         fileCsv = TestFileUtil.createNewFile("largefileCsv.csv");
     }
 
     @Test
-    public void t01Read()  throws Exception{
-        Thread.sleep(10*1000L);
+    public void t01Read() throws Exception {
         long start = System.currentTimeMillis();
         EasyExcel.read(TestFileUtil.getPath() + "large" + File.separator + "large07.xlsx", LargeData.class,
             new LargeDataListener()).headRowNumber(2).sheet().doRead();
@@ -86,9 +87,16 @@ public class LargeDataTest {
 
     @Test
     public void t04Write() throws Exception {
-        long start = System.currentTimeMillis();
-        ExcelWriter excelWriter = EasyExcel.write(fileWrite07, LargeData.class).build();
+        ExcelWriter excelWriter = EasyExcel.write(fileWriteTemp07, LargeData.class).build();
         WriteSheet writeSheet = EasyExcel.writerSheet().build();
+        for (int j = 0; j < 2; j++) {
+            excelWriter.write(data(), writeSheet);
+        }
+        excelWriter.finish();
+
+        long start = System.currentTimeMillis();
+        excelWriter = EasyExcel.write(fileWrite07, LargeData.class).build();
+        writeSheet = EasyExcel.writerSheet().build();
         for (int j = 0; j < 100; j++) {
             excelWriter.write(data(), writeSheet);
             LOGGER.info("{} write success.", j);
@@ -116,7 +124,8 @@ public class LargeDataTest {
         }
         long costPoi = System.currentTimeMillis() - start;
         LOGGER.info("poi write cost:{}", System.currentTimeMillis() - start);
-        Assert.assertTrue(costPoi * 3 > cost);
+        LOGGER.info("{} vs {}", cost, costPoi);
+        Assert.assertTrue(costPoi * 2 > cost);
     }
 
     private List<LargeData> data() {

update.md

@@ -1,5 +1,5 @@
 # 3.0.2
-* 修复写入的性能问题
+* 大幅提升读写性能
 
 # 3.0.1
 * 升级到正式版