iterator over sheet data in XSSFReader returns sheets in logical order, i.e. as they are defined in workbook.xml (was in physical order, as they were stored in the relationship table)

git-svn-id: https://svn.apache.org/repos/asf/poi/branches/ooxml@700821 13f79535-47bb-0310-9956-ffa450edef68

src/ooxml/java/org/apache/poi/xssf/eventusermodel/XSSFReader.java

@@ -18,20 +18,24 @@ package org.apache.poi.xssf.eventusermodel;
 
 import java.io.IOException;
 import java.io.InputStream;
-import java.util.Iterator;
+import java.util.*;
 
 import org.apache.poi.xssf.model.SharedStringsTable;
 import org.apache.poi.xssf.model.StylesTable;
 import org.apache.poi.xssf.usermodel.XSSFRelation;
+import org.apache.poi.POIXMLException;
+import org.apache.xmlbeans.XmlException;
 import org.openxml4j.exceptions.InvalidFormatException;
 import org.openxml4j.exceptions.OpenXML4JException;
 import org.openxml4j.opc.Package;
 import org.openxml4j.opc.PackagePart;
 import org.openxml4j.opc.PackagePartName;
 import org.openxml4j.opc.PackageRelationship;
-import org.openxml4j.opc.PackageRelationshipCollection;
 import org.openxml4j.opc.PackageRelationshipTypes;
 import org.openxml4j.opc.PackagingURIHelper;
+import org.openxmlformats.schemas.spreadsheetml.x2006.main.CTWorkbook;
+import org.openxmlformats.schemas.spreadsheetml.x2006.main.WorkbookDocument;
+import org.openxmlformats.schemas.spreadsheetml.x2006.main.CTSheet;
 
 /**
  * This class makes it easy to get at individual parts
@@ -129,37 +133,96 @@ public class XSSFReader {
      *  InputStreams when done with each one.
      */
     public Iterator<InputStream> getSheetsData() throws IOException, InvalidFormatException {
-		return new SheetDataIterator();
+        return new SheetIterator(workbookPart);
     }
 
-	private class SheetDataIterator implements Iterator<InputStream> {
+    /**
-		private Iterator<PackageRelationship> sheetRels;
+     * Iterator over sheet data.
-		private SheetDataIterator() throws IOException, InvalidFormatException {
+     */
-			// Find all the sheets
+    public static class SheetIterator implements Iterator<InputStream> {
-			PackageRelationshipCollection sheets =
-				workbookPart.getRelationshipsByType(
-						XSSFRelation.WORKSHEET.getRelation()
-			);
-			sheetRels = sheets.iterator();
-		}
 
-		public boolean hasNext() {
+        /**
-			return sheetRels.hasNext();
+         *  Maps relId and the corresponding PackagePart
-		}
+         */
+        private Map<String, PackagePart> sheetMap;
 
-		public InputStream next() {
+        /**
-			PackageRelationship sheet = sheetRels.next();
+         * Current CTSheet bean
+         */
+        private CTSheet ctSheet;
+
+        /**
+         * Iterator over CTSheet objects, returns sheets in <tt>logical</tt> order.
+         * We can't rely on the Ooxml4J's relationship iterator because it returns objects in physical order,
+         * i.e. as they are stored in the underlying package
+         */
+        private Iterator<CTSheet> sheetIterator;
+
+        /**
+         * Construct a new SheetIterator
+         *
+         * @param wb package part holding workbook.xml
+         */
+        private SheetIterator(PackagePart wb) throws IOException {
+
+            /**
+             * The order of sheets is defined by the order of CTSheet elements in workbook.xml
+             */
             try {
-		        PackagePartName relName = PackagingURIHelper.createPartName(sheet.getTargetURI());
+                //step 1. Map sheet's relationship Id and the corresponding PackagePart
-				PackagePart sheetPkg = pkg.getPart(relName);
+                sheetMap = new HashMap<String, PackagePart>();
+                for(PackageRelationship rel : wb.getRelationships()){
+                    if(rel.getRelationshipType().equals(XSSFRelation.WORKSHEET.getRelation())){
+                        PackagePartName relName = PackagingURIHelper.createPartName(rel.getTargetURI());
+                        sheetMap.put(rel.getId(), wb.getPackage().getPart(relName));
+                    }
+                }
+                //step 2. Read array of CTSheet elements, wrap it in a ArayList and construct an iterator
+                //Note, using XMLBeans might be expensive, consider refactoring to use SAX or a plain regexp search
+                CTWorkbook wbBean = WorkbookDocument.Factory.parse(wb.getInputStream()).getWorkbook();
+                sheetIterator = Arrays.asList(wbBean.getSheets().getSheetArray()).iterator();
+            } catch (InvalidFormatException e){
+                throw new POIXMLException(e);
+            } catch (XmlException e){
+                throw new POIXMLException(e);
+            }
+        }
+
+        /**
+         * Returns <tt>true</tt> if the iteration has more elements.
+         *
+         * @return <tt>true</tt> if the iterator has more elements.
+         */
+        public boolean hasNext() {
+            return sheetIterator.hasNext();
+        }
+
+        /**
+         * Returns input stream of the next sheet in the iteration
+         *
+         * @return input stream of the next sheet in the iteration
+         */
+        public InputStream next() {
+            ctSheet = sheetIterator.next();
+
+            String sheetId = ctSheet.getId();
+            try {
+                PackagePart sheetPkg = sheetMap.get(sheetId);
                 return sheetPkg.getInputStream();
             } catch(IOException e) {
-				throw new RuntimeException(e);
+                throw new POIXMLException(e);
-			} catch(InvalidFormatException ife) {
-				throw new RuntimeException(ife);
             }
         }
 
+        /**
+         * Returns name of the current sheet
+         *
+         * @return name of the current sheet
+         */
+        public String getSheetName() {
+            return ctSheet.getName();
+        }
+
         public void remove() {
             throw new IllegalStateException("Not supported");
         }

src/ooxml/testcases/org/apache/poi/xssf/eventusermodel/TestXSSFReader.java

@@ -109,4 +109,30 @@ public class TestXSSFReader extends TestCase {
     	}
     	assertEquals(3, count);
     }
+
+    /**
+     * Check that the sheet iterator returns sheets in the logical order
+     * (as they are defined in the workbook.xml)
+     */
+    public void testOrderOfSheets() throws Exception {
+        File f = new File(dirName, "reordered_sheets.xlsx");
+        Package pkg = Package.open(f.toString());
+
+        XSSFReader r = new XSSFReader(pkg);
+
+        String[] sheetNames = {"Sheet4", "Sheet2", "Sheet3", "Sheet1"};
+        XSSFReader.SheetIterator it = (XSSFReader.SheetIterator)r.getSheetsData();
+
+        int count = 0;
+        while(it.hasNext()) {
+            InputStream inp = it.next();
+            assertNotNull(inp);
+            inp.close();
+
+            assertEquals(sheetNames[count], it.getSheetName());
+            count++;
+        }
+        assertEquals(4, count);
+
+    }
 }