Cell Style optimisations too

git-svn-id: https://svn.apache.org/repos/asf/poi/trunk@677057 13f79535-47bb-0310-9956-ffa450edef68

src/java/org/apache/poi/hssf/model/Workbook.java

@@ -714,6 +714,18 @@ public class Workbook implements Model
         return retval;
     }
     
+    /**
+     * Removes the given ExtendedFormatRecord record from the
+     *  file's list. This will make all 
+     *  subsequent font indicies drop by one,
+     *  so you'll need to update those yourself!
+     */
+    public void removeExFormatRecord(ExtendedFormatRecord rec) {
+    	records.remove(rec); // this updates XfPos for us
+        numxfs--;
+    }
+
+
     /**
      * creates a new Cell-type Extneded Format Record and adds it to the end of
      *  ExtendedFormatRecords collection

src/java/org/apache/poi/hssf/usermodel/HSSFCell.java

@@ -927,7 +927,6 @@ public class HSSFCell
      * object.
      * @see org.apache.poi.hssf.usermodel.HSSFWorkbook#getCellStyleAt(short)
      */
-
     public HSSFCellStyle getCellStyle()
     {
       short styleIndex=record.getXFIndex();

src/java/org/apache/poi/hssf/usermodel/HSSFOptimiser.java

@@ -173,6 +173,89 @@ public class HSSFOptimiser {
 	 * @param workbook The workbook in which to optimise the cell styles
 	 */
 	public static void optimiseCellStyles(HSSFWorkbook workbook) {
+		// Where each style has ended up, and if we need to
+		//  delete the record for it. Start off with no change
+		short[] newPos = 
+			new short[workbook.getWorkbook().getNumExFormats()];
+		boolean[] zapRecords = new boolean[newPos.length];
+		for(int i=0; i<newPos.length; i++) {
+			newPos[i] = (short)i;
+			zapRecords[i] = false;
+		}
 		
+		// Get each style record, so we can do deletes
+		//  without getting confused
+		ExtendedFormatRecord[] xfrs = new ExtendedFormatRecord[newPos.length]; 
+		for(int i=0; i<newPos.length; i++) {
+			xfrs[i] = workbook.getWorkbook().getExFormatAt(i);
+		}
+		
+		// Loop over each style, seeing if it is the same
+		//  as an earlier one. If it is, point users of the
+		//  later duplicate copy to the earlier one, and 
+		//  mark the later one as needing deleting
+		// Only work on user added ones, which come after 20
+		for(int i=21; i<newPos.length; i++) {
+			// Check this one for being a duplicate
+			//  of an earlier one
+			int earlierDuplicate = -1;
+			for(int j=0; j<i && earlierDuplicate == -1; j++) {
+				ExtendedFormatRecord xfCheck = workbook.getWorkbook().getExFormatAt(j);
+				if(xfCheck.equals(xfrs[i])) {
+					earlierDuplicate = j;
+				}
+			}
+			
+			// If we got a duplicate, mark it as such
+			if(earlierDuplicate != -1) {
+				newPos[i] = (short)earlierDuplicate;
+				zapRecords[i] = true;
+			}
+		}
+		
+		// Update the new positions based on
+		//  deletes that have occurred between
+		//  the start and them
+		// Only work on user added ones, which come after 20
+		for(int i=21; i<newPos.length; i++) {
+			// Find the number deleted to that
+			//  point, and adjust
+			short preDeletePos = newPos[i];
+			short newPosition = preDeletePos;
+			for(int j=0; j<preDeletePos; j++) {
+				if(zapRecords[j]) newPosition--;
+			}
+			
+			// Update the new position
+			newPos[i] = newPosition;
+		}
+		
+		// Zap the un-needed user style records
+		for(int i=21; i<newPos.length; i++) {
+			if(zapRecords[i]) {
+				workbook.getWorkbook().removeExFormatRecord(
+						xfrs[i]
+				);
+			}
+		}
+		
+		// Finally, update the cells to point at
+		//  their new extended format records
+		for(int sheetNum=0; sheetNum<workbook.getNumberOfSheets(); sheetNum++) {
+			HSSFSheet s = workbook.getSheetAt(sheetNum);
+			Iterator rIt = s.rowIterator();
+			while(rIt.hasNext()) {
+				HSSFRow row = (HSSFRow)rIt.next();
+				Iterator cIt = row.cellIterator();
+				while(cIt.hasNext()) {
+					HSSFCell cell = (HSSFCell)cIt.next();
+					short oldXf = cell.getCellValueRecord().getXFIndex();
+					HSSFCellStyle newStyle = workbook.getCellStyleAt(
+							newPos[oldXf]
+					);
+					cell.setCellStyle(newStyle);
+				}
+			}
+		}
 	}
 }

src/testcases/org/apache/poi/hssf/usermodel/TestHSSFOptimiser.java

@@ -15,6 +15,8 @@
 ==================================================================== */
 package org.apache.poi.hssf.usermodel;
 
+import org.apache.poi.hssf.model.Workbook;
+
 import junit.framework.TestCase;
 
 public class TestHSSFOptimiser extends TestCase {
@@ -38,7 +40,7 @@ public class TestHSSFOptimiser extends TestCase {
 		assertEquals(f, s.getFont(wb));
 		
 		// Optimise styles
-//		HSSFOptimiser.optimiseCellStyles(wb);
+		HSSFOptimiser.optimiseCellStyles(wb);
 		
 		assertEquals(5, wb.getNumberOfFonts());
 		assertEquals(22, wb.getNumCellStyles());
@@ -77,6 +79,8 @@ public class TestHSSFOptimiser extends TestCase {
 		
 		
 		// Use all three of the four in cell styles
+		assertEquals(21, wb.getNumCellStyles());
+		
 		HSSFCellStyle cs1 = wb.createCellStyle();
 		cs1.setFont(f1);
 		assertEquals(5, cs1.getFontIndex());
@@ -93,6 +97,8 @@ public class TestHSSFOptimiser extends TestCase {
 		cs4.setFont(f6);
 		assertEquals(10, cs4.getFontIndex());
 		
+		assertEquals(25, wb.getNumCellStyles());
+		
 		
 		// And three in rich text
 		HSSFSheet s = wb.createSheet();
@@ -144,4 +150,91 @@ public class TestHSSFOptimiser extends TestCase {
 		assertEquals(8, r.getCell(1).getRichStringCellValue().getFontAtIndex(6));
 		assertEquals(8, r.getCell(1).getRichStringCellValue().getFontAtIndex(7));
 	}
+	
+	public void testOptimiseStyles() throws Exception {
+		HSSFWorkbook wb = new HSSFWorkbook();
+
+		// Two fonts
+		assertEquals(4, wb.getNumberOfFonts());
+		
+		HSSFFont f1 = wb.createFont();
+		f1.setFontHeight((short)11);
+		f1.setFontName("Testing");
+		
+		HSSFFont f2 = wb.createFont();
+		f2.setFontHeight((short)22);
+		f2.setFontName("Also Testing");
+
+		assertEquals(6, wb.getNumberOfFonts());
+	
+		
+		// Several styles
+		assertEquals(21, wb.getNumCellStyles());
+		
+		HSSFCellStyle cs1 = wb.createCellStyle();
+		cs1.setFont(f1);
+		
+		HSSFCellStyle cs2 = wb.createCellStyle();
+		cs2.setFont(f2);
+		
+		HSSFCellStyle cs3 = wb.createCellStyle();
+		cs3.setFont(f1);
+		
+		HSSFCellStyle cs4 = wb.createCellStyle();
+		cs4.setFont(f1);
+		cs4.setAlignment((short)22);
+		
+		HSSFCellStyle cs5 = wb.createCellStyle();
+		cs5.setFont(f2);
+		cs5.setAlignment((short)111);
+		
+		HSSFCellStyle cs6 = wb.createCellStyle();
+		cs6.setFont(f2);
+		
+		assertEquals(27, wb.getNumCellStyles());
+		
+		
+		// Use them
+		HSSFSheet s = wb.createSheet();
+		HSSFRow r = s.createRow(0);
+		
+		r.createCell((short)0).setCellStyle(cs1);
+		r.createCell((short)1).setCellStyle(cs2);
+		r.createCell((short)2).setCellStyle(cs3);
+		r.createCell((short)3).setCellStyle(cs4);
+		r.createCell((short)4).setCellStyle(cs5);
+		r.createCell((short)5).setCellStyle(cs6);
+		r.createCell((short)6).setCellStyle(cs1);
+		r.createCell((short)7).setCellStyle(cs2);
+		
+		assertEquals(21, r.getCell(0).getCellValueRecord().getXFIndex());
+		assertEquals(26, r.getCell(5).getCellValueRecord().getXFIndex());
+		assertEquals(21, r.getCell(6).getCellValueRecord().getXFIndex());
+		
+		
+		// Optimise
+		HSSFOptimiser.optimiseCellStyles(wb);
+		
+		
+		// Check
+		assertEquals(6, wb.getNumberOfFonts());
+		assertEquals(25, wb.getNumCellStyles());
+		
+		// cs1 -> 21
+		assertEquals(21, r.getCell(0).getCellValueRecord().getXFIndex());
+		// cs2 -> 22
+		assertEquals(22, r.getCell(1).getCellValueRecord().getXFIndex());
+		// cs3 = cs1 -> 21
+		assertEquals(21, r.getCell(2).getCellValueRecord().getXFIndex());
+		// cs4 --> 24 -> 23
+		assertEquals(23, r.getCell(3).getCellValueRecord().getXFIndex());
+		// cs5 --> 25 -> 24
+		assertEquals(24, r.getCell(4).getCellValueRecord().getXFIndex());
+		// cs6 = cs2 -> 22
+		assertEquals(22, r.getCell(5).getCellValueRecord().getXFIndex());
+		// cs1 -> 21
+		assertEquals(21, r.getCell(6).getCellValueRecord().getXFIndex());
+		// cs2 -> 22
+		assertEquals(22, r.getCell(7).getCellValueRecord().getXFIndex());
+	}
 }