More on converting the excel extractor to the new code

git-svn-id: https://svn.apache.org/repos/asf/poi/branches/ooxml@635030 13f79535-47bb-0310-9956-ffa450edef68

src/ooxml/java/org/apache/poi/xssf/extractor/XSSFExcelExtractor.java

@@ -18,12 +18,14 @@ package org.apache.poi.xssf.extractor;
 
 import java.io.File;
 import java.io.IOException;
+import java.util.Iterator;
 
 import org.apache.poi.POIXMLTextExtractor;
 import org.apache.poi.ss.usermodel.Cell;
 import org.apache.poi.ss.usermodel.Row;
 import org.apache.poi.ss.usermodel.Sheet;
 import org.apache.poi.ss.usermodel.Workbook;
+import org.apache.poi.xssf.usermodel.XSSFCell;
 import org.apache.poi.xssf.usermodel.XSSFSheet;
 import org.apache.poi.xssf.usermodel.XSSFWorkbook;
 import org.apache.xmlbeans.XmlException;
@@ -92,17 +94,21 @@ public class XSSFExcelExtractor extends POIXMLTextExtractor {
 			
 			for (Object rawR : sheet) {
 				Row row = (Row)rawR;
-				for (Object rawC: row) {
-					Cell cell = (Cell)rawC;
+				for(Iterator ri = row.cellIterator(); ri.hasNext();) {
+					Cell cell = (Cell)ri.next();
 					
 					// Is it a formula one?
 					if(cell.getCellType() == Cell.CELL_TYPE_FORMULA && formulasNotResults) {
 						text.append(cell.getCellFormula());
+					} else if(cell.getCellType() == Cell.CELL_TYPE_STRING) {
+						text.append(cell.getRichStringCellValue().getString());
 					} else {
-						text.append(cell.toString());
+						XSSFCell xc = (XSSFCell)cell;
+						text.append(xc.getRawValue());
 					}
 					
-					text.append(",");
+					if(ri.hasNext())
+						text.append("\t");
 				}
 				text.append("\n");
 			}

src/ooxml/java/org/apache/poi/xssf/usermodel/XSSFCell.java

@@ -78,7 +78,7 @@ public class XSSFCell implements Cell {
     }
 
     public String getCellFormula() {
-        if (STCellType.STR != cell.getT()) { 
+        if(this.cell.getF() == null) { 
             throw new NumberFormatException("You cannot get a formula from a non-formula cell");
         }
         return this.cell.getF().getStringValue();
@@ -94,6 +94,12 @@ public class XSSFCell implements Cell {
     }
 
     public int getCellType() {
+    	// Detecting formulas is quite pesky,
+    	//  as they don't get their type set
+    	if(this.cell.getF() != null) {
+    		return CELL_TYPE_FORMULA;
+    	}
+    	
         switch (this.cell.getT().intValue()) {
         case STCellType.INT_B:
             return CELL_TYPE_BOOLEAN;
@@ -289,6 +295,13 @@ public class XSSFCell implements Cell {
     public String toString() {
         return "[" + this.row.getRowNum() + "," + this.getCellNum() + "] " + this.cell.getV();
     }
+    
+    /**
+     * Returns the raw, underlying ooxml value for the cell
+     */
+    public String getRawValue() {
+    	return this.cell.getV();
+    }
 
     /**
      * @throws RuntimeException if the bounds are exceeded.

src/ooxml/java/org/apache/poi/xssf/usermodel/XSSFRichTextString.java

@@ -79,6 +79,9 @@ public class XSSFRichTextString implements RichTextString {
     public String getString() {
         return string;
     }
+    public String toString() {
+    	return string;
+    }
 
     public int length() {
         return string.length();

src/ooxml/testcases/org/apache/poi/xssf/extractor/TestXSSFExcelExtractor.java

@@ -35,59 +35,56 @@ public class TestXSSFExcelExtractor extends TestCase {
 	/**
 	 * A very simple file
 	 */
-	private XSSFWorkbook xmlA;
-	private File fileA;
+	private File xmlA;
 	/**
 	 * A fairly complex file
 	 */
-	private XSSFWorkbook xmlB;
+	private File xmlB;
 	
 	/**
 	 * A fairly simple file - ooxml
 	 */
-	private XSSFWorkbook simpleXLSX; 
+	private File simpleXLSX; 
 	/**
 	 * A fairly simple file - ole2
 	 */
-	private HSSFWorkbook simpleXLS;
+	private File simpleXLS;
 
 	protected void setUp() throws Exception {
 		super.setUp();
 		
-		fileA = new File(
+		xmlA = new File(
 				System.getProperty("HSSF.testdata.path") +
 				File.separator + "sample.xlsx"
 		);
-		File fileB = new File(
+		assertTrue(xmlA.exists());
+		xmlB = new File(
 				System.getProperty("HSSF.testdata.path") +
 				File.separator + "AverageTaxRates.xlsx"
 		);
+		assertTrue(xmlB.exists());
 		
-		File fileSOOXML = new File(
+		simpleXLSX = new File(
 				System.getProperty("HSSF.testdata.path") +
 				File.separator + "SampleSS.xlsx"
 		);
-		File fileSOLE2 = new File(
+		simpleXLS = new File(
 				System.getProperty("HSSF.testdata.path") +
 				File.separator + "SampleSS.xls"
 		);
-		
-		xmlA = new XSSFWorkbook(fileA.toString());
-		xmlB = new XSSFWorkbook(fileB.toString());
-		
-		simpleXLSX = new XSSFWorkbook(fileSOOXML.toString());
-		simpleXLS = new HSSFWorkbook(new FileInputStream(fileSOLE2));
+		assertTrue(simpleXLS.exists());
+		assertTrue(simpleXLSX.exists());
 	}
 
 	/**
 	 * Get text out of the simple file
 	 */
 	public void testGetSimpleText() throws Exception {
-		new XSSFExcelExtractor(fileA.toString());
-		new XSSFExcelExtractor(xmlA);
+		new XSSFExcelExtractor(xmlA.toString());
+		new XSSFExcelExtractor(new XSSFWorkbook(xmlA.toString()));
 		
 		XSSFExcelExtractor extractor = 
-			new XSSFExcelExtractor(xmlA);
+			new XSSFExcelExtractor(xmlA.toString());
 		extractor.getText();
 		
 		String text = extractor.getText();
@@ -110,8 +107,7 @@ public class TestXSSFExcelExtractor extends TestCase {
 				"adipiscing\t777\n" +
 				"elit\t888\n" +
 				"Nunc\t999\n" +
-				"at\t4995\n" +
-				"\n\n", text);
+				"at\t4995\n", text);
 		
 		// Now get formulas not their values
 		extractor.setFormulasNotResults(true);
@@ -126,8 +122,7 @@ public class TestXSSFExcelExtractor extends TestCase {
 				"adipiscing\t777\n" +
 				"elit\t888\n" +
 				"Nunc\t999\n" +
-				"at\tSUM(B1:B9)\n" +
-				"\n\n", text);
+				"at\tSUM(B1:B9)\n", text);
 		
 		// With sheet names too
 		extractor.setIncludeSheetNames(true);
@@ -143,17 +138,17 @@ public class TestXSSFExcelExtractor extends TestCase {
 				"adipiscing\t777\n" +
 				"elit\t888\n" +
 				"Nunc\t999\n" +
-				"at\tSUM(B1:B9)\n\n" +
-				"Sheet2\n\n" +
+				"at\tSUM(B1:B9)\n" +
+				"Sheet2\n" +
 				"Sheet3\n"
 				, text);
 	}
 	
 	public void testGetComplexText() throws Exception {
-		new XSSFExcelExtractor(xmlB);
+		new XSSFExcelExtractor(xmlB.toString());
 		
 		XSSFExcelExtractor extractor = 
-			new XSSFExcelExtractor(xmlB);
+			new XSSFExcelExtractor(new XSSFWorkbook(xmlB.toString()));
 		extractor.getText();
 		
 		String text = extractor.getText();
@@ -174,9 +169,10 @@ public class TestXSSFExcelExtractor extends TestCase {
 	 */
 	public void testComparedToOLE2() throws Exception {
 		XSSFExcelExtractor ooxmlExtractor =
-			new XSSFExcelExtractor(simpleXLSX);
+			new XSSFExcelExtractor(simpleXLSX.toString());
 		ExcelExtractor ole2Extractor =
-			new ExcelExtractor(simpleXLS);
+			new ExcelExtractor(new HSSFWorkbook(
+					new FileInputStream(simpleXLS)));
 		
 		POITextExtractor[] extractors =
 			new POITextExtractor[] { ooxmlExtractor, ole2Extractor };