Have HSSFOptimiser also remove un-used cell styles, in addition to duplicates

git-svn-id: https://svn.apache.org/repos/asf/poi/trunk@1391891 13f79535-47bb-0310-9956-ffa450edef68

src/documentation/content/xdocs/status.xml

@@ -34,6 +34,7 @@
 
     <changes>
         <release version="3.9-beta1" date="2012-??-??">
+          <action dev="poi-developers" type="add">HSSFOptimiser will now also tidy away un-used cell styles, in addition to duplicate styles</action>
           <action dev="poi-developers" type="fix">53493 - Fixed memory and temporary file leak in SXSSF </action>
           <action dev="poi-developers" type="fix">53780 - Fixed memory and temporary file leak in SXSSF </action>
           <action dev="poi-developers" type="fix">53380 - ArrayIndexOutOfBounds Excetion parsing word 97 document. </action>

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

@@ -165,7 +165,7 @@ public class HSSFOptimiser {
 	
 	/**
 	 * Goes through the Wokrbook, optimising the cell styles
-	 *  by removing duplicate ones.
+	 *  by removing duplicate ones, and ones that aren't used.
 	 * For best results, optimise the fonts via a call to
 	 *  {@link #optimiseFonts(HSSFWorkbook)} first.
 	 * @param workbook The workbook in which to optimise the cell styles
@@ -175,8 +175,10 @@ public class HSSFOptimiser {
 		//  delete the record for it. Start off with no change
 		short[] newPos = 
 			new short[workbook.getWorkbook().getNumExFormats()];
+		boolean[] isUsed = new boolean[newPos.length];
 		boolean[] zapRecords = new boolean[newPos.length];
 		for(int i=0; i<newPos.length; i++) {
+         isUsed[i] = false;
 			newPos[i] = (short)i;
 			zapRecords[i] = false;
 		}
@@ -211,6 +213,27 @@ public class HSSFOptimiser {
 			}
 		}
 		
+		// Loop over all the cells in the file, and identify any user defined
+		//  styles aren't actually being used (don't touch built-in ones)
+      for(int sheetNum=0; sheetNum<workbook.getNumberOfSheets(); sheetNum++) {
+         HSSFSheet s = workbook.getSheetAt(sheetNum);
+         for (Row row : s) {
+            for (Cell cellI : row) {
+               HSSFCell cell = (HSSFCell)cellI;
+               short oldXf = cell.getCellValueRecord().getXFIndex();
+               isUsed[oldXf] = true;
+            }
+         }
+      }
+      // Mark any that aren't used as needing zapping
+      for (int i=21; i<isUsed.length; i++) {
+         if (! isUsed[i]) {
+            // Un-used style, can be removed
+            zapRecords[i] = true;
+            newPos[i] = 0;
+         }
+      }
+		
 		// Update the new positions based on
 		//  deletes that have occurred between
 		//  the start and them
@@ -237,14 +260,14 @@ public class HSSFOptimiser {
 			}
 		}
 		
-		// Finally, update the cells to point at
+		// Finally, update the cells to point at their new extended format records
-		//  their new extended format records
 		for(int sheetNum=0; sheetNum<workbook.getNumberOfSheets(); sheetNum++) {
 			HSSFSheet s = workbook.getSheetAt(sheetNum);
 			for (Row row : s) {
 			   for (Cell cellI : row) {
 					HSSFCell cell = (HSSFCell)cellI;
 					short oldXf = cell.getCellValueRecord().getXFIndex();
+					
 					HSSFCellStyle newStyle = workbook.getCellStyleAt(
 							newPos[oldXf]
 					);

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

@@ -22,11 +22,21 @@ public final class TestHSSFOptimiser extends TestCase {
 	public void testDoesNoHarmIfNothingToDo() {
 		HSSFWorkbook wb = new HSSFWorkbook();
 
+		// New files start with 4 built in fonts, and 21 built in styles
+      assertEquals(4, wb.getNumberOfFonts());
+      assertEquals(21, wb.getNumCellStyles());
+
+      // Create a test font and style, and use them
       HSSFFont f = wb.createFont();
 		f.setFontName("Testing");
 		HSSFCellStyle s = wb.createCellStyle();
 		s.setFont(f);
 		
+		HSSFSheet sheet = wb.createSheet();
+		HSSFRow row = sheet.createRow(0);
+		row.createCell(0).setCellStyle(s);
+
+		// Should have one more than the default of each
 		assertEquals(5, wb.getNumberOfFonts());
 		assertEquals(22, wb.getNumCellStyles());
 
@@ -226,5 +236,37 @@ public final class TestHSSFOptimiser extends TestCase {
 		assertEquals(21, r.getCell(6).getCellValueRecord().getXFIndex());
 		// cs2 -> 22
 		assertEquals(22, r.getCell(7).getCellValueRecord().getXFIndex());
+		
+		
+		// Add a new duplicate, and two that aren't used
+      HSSFCellStyle csD = wb.createCellStyle();
+      csD.setFont(f1);
+      r.createCell(8).setCellStyle(csD);
+      
+      HSSFFont f3 = wb.createFont();
+      f3.setFontHeight((short) 23);
+      f3.setFontName("Testing 3");
+      HSSFFont f4 = wb.createFont();
+      f4.setFontHeight((short) 24);
+      f4.setFontName("Testing 4");
+
+      HSSFCellStyle csU1 = wb.createCellStyle();
+      csU1.setFont(f3);
+      HSSFCellStyle csU2 = wb.createCellStyle();
+      csU2.setFont(f4);
+      
+      // Check before the optimise
+      assertEquals(8, wb.getNumberOfFonts());
+      assertEquals(28, wb.getNumCellStyles());
+      
+      // Optimise, should remove the two un-used ones and the one duplicate
+      HSSFOptimiser.optimiseCellStyles(wb);
+      
+      // Check
+      assertEquals(8, wb.getNumberOfFonts());
+      assertEquals(25, wb.getNumCellStyles());
+      
+      // csD -> cs1 -> 21
+      assertEquals(21, r.getCell(8).getCellValueRecord().getXFIndex());
 	}
 }