优化07版超大文件读取参数配置

src/main/java/com/alibaba/excel/analysis/v07/XlsxSaxAnalyser.java

@@ -103,7 +103,7 @@ public class XlsxSaxAnalyser implements ExcelExecutor {
         if (size < 0) {
             size = sharedStringsTablePackagePart.getInputStream().available();
         }
-        if (size < USE_MAP_CACHE_SIZE) {
+        if (size < readWorkbookHolder.getSharedStringsUseEhcacheSize()) {
             if (LOGGER.isDebugEnabled()) {
                 LOGGER.debug("Use map cache.size:{}", size);
             }

src/main/java/com/alibaba/excel/cache/Ehcache.java

@@ -32,9 +32,7 @@ public class Ehcache implements ReadCache {
     private static final Logger LOGGER = LoggerFactory.getLogger(Ehcache.class);
 
     private static final int BATCH_COUNT = 1000;
-    private static final int CHECK_INTERVAL = 500;
-    private static final int MAX_CACHE_ACTIVATE = 10;
-
+    private static final int CHECK_INTERVAL = 1000;
     private static final String CACHE = "cache";
     private static final String DATA_SEPARATOR = "@";
     private static final String KEY_VALUE_SEPARATOR = "!";
@@ -50,6 +48,10 @@ public class Ehcache implements ReadCache {
      * Key index
      */
     private int index = 0;
+    /**
+     * Maximum number of cache activation
+     */
+    private int maxCacheActivate = 100;
     private StringBuilder data = new StringBuilder();
     private CacheManager cacheManager;
     /**
@@ -80,6 +82,12 @@ public class Ehcache implements ReadCache {
      */
     private int cacheMiss = 0;
 
+    public Ehcache() {}
+
+    public Ehcache(int maxCacheActivate) {
+        this.maxCacheActivate = maxCacheActivate;
+    }
+
     @Override
     public void init(AnalysisContext analysisContext) {
         File readTempFile = analysisContext.readWorkbookHolder().getTempFile();
@@ -166,7 +174,7 @@ public class Ehcache implements ReadCache {
     }
 
     private void checkClear() {
-        if (countList.size() > MAX_CACHE_ACTIVATE) {
+        if (countList.size() > maxCacheActivate) {
             Integer route = countList.getFirst();
             countList.removeFirst();
             cacheMap.remove(route);

src/main/java/com/alibaba/excel/write/metadata/holder/WriteWorkbookHolder.java

@@ -10,8 +10,6 @@ import java.util.HashMap;
 import java.util.Map;
 
 import org.apache.poi.ss.usermodel.Workbook;
-import org.slf4j.Logger;
-import org.slf4j.LoggerFactory;
 
 import com.alibaba.excel.enums.HolderEnum;
 import com.alibaba.excel.exception.ExcelGenerateException;

src/test/java/com/alibaba/easyexcel/test/temp/large/LargeData.java

+package com.alibaba.easyexcel.test.temp.large;
+
+import lombok.Data;
+
+/**
+ * @author Jiaju Zhuang
+ */
+@Data
+public class LargeData {
+
+    private String str1;
+
+    private String str2;
+
+    private String str3;
+
+    private String str4;
+
+    private String str5;
+
+    private String str6;
+
+    private String str7;
+
+    private String str8;
+
+    private String str9;
+
+    private String str10;
+
+    private String str11;
+
+    private String str12;
+
+    private String str13;
+
+    private String str14;
+
+    private String str15;
+
+    private String str16;
+
+    private String str17;
+
+    private String str18;
+
+    private String str19;
+
+    private String str20;
+
+    private String str21;
+
+    private String str22;
+
+    private String str23;
+
+    private String str24;
+
+    private String str25;
+}

src/test/java/com/alibaba/easyexcel/test/temp/large/LargeDataListener.java

+package com.alibaba.easyexcel.test.temp.large;
+
+import org.junit.Assert;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import com.alibaba.excel.context.AnalysisContext;
+import com.alibaba.excel.event.AnalysisEventListener;
+import com.alibaba.fastjson.JSON;
+
+/**
+ * @author Jiaju Zhuang
+ */
+public class LargeDataListener extends AnalysisEventListener<LargeData> {
+    private static final Logger LOGGER = LoggerFactory.getLogger(LargeDataListener.class);
+    private int count = 0;
+
+    @Override
+    public void invoke(LargeData data, AnalysisContext context) {
+        if (count == 0) {
+            LOGGER.info("First row:{}", JSON.toJSONString(data));
+        }
+        count++;
+        if (count % 100000 == 0) {
+            LOGGER.info("Already read:{}", count);
+        }
+    }
+
+    @Override
+    public void doAfterAllAnalysed(AnalysisContext context) {
+        LOGGER.info("Large row count:{}", count);
+    }
+}

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

+package com.alibaba.easyexcel.test.temp.large;
+
+import java.io.FileInputStream;
+
+import org.junit.Test;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import com.alibaba.excel.EasyExcel;
+import com.alibaba.excel.cache.MapCache;
+
+/**
+ *
+ * @author Jiaju Zhuang
+ */
+public class LargeDataTest {
+    private static final Logger LOGGER = LoggerFactory.getLogger(LargeDataTest.class);
+
+    @Test
+    public void read() throws Exception {
+        long start = System.currentTimeMillis();
+        EasyExcel.read(new FileInputStream("D:\\test\\MRP生产视图(1).xlsx"), LargeData.class, new LargeDataListener()).readCache(new eh)
+            .headRowNumber(2).sheet().doRead();
+        LOGGER.info("Large data total time spent:{}", System.currentTimeMillis() - start);
+    }
+}

update.md

+# 2.0.5
+* 优化07版超大文件读取参数配置
+* 支持自己设置超大文件读取参数
+
 # 2.0.4
 * 修复07版整个excel仅存在数字时会出现的NPE
 * 修复03版 用String接收电话会出现科学计数法的问题