Examples of ExcelExtractor

org.apache.poi.hssf.extractor.ExcelExtractor
A text extractor for Excel files. Returns the textual content of the file, suitable for indexing by something like Lucene, but not really intended for display to the user. To turn an excel file into a CSV or similar, then see the XLS2CSVmra example @see org.apache.poi.hssf.eventusermodel.examples.XLS2CSVmra

Examples of org.apache.poi.hssf.extractor.ExcelExtractor

    //  out from
    for(Iterator entries = poifsDir.getEntries(); entries.hasNext(); ) {
      Entry entry = (Entry)entries.next();
      
      if(entry.getName().equals("Workbook")) {
        return new ExcelExtractor(poifsDir, fs);
      }
      if(entry.getName().equals("WordDocument")) {
        return new WordExtractor(poifsDir, fs);
      }
      if(entry.getName().equals("PowerPoint Document")) {

View Full Code Here

Examples of org.apache.poi.hssf.extractor.ExcelExtractor

    //  out from
    for(Iterator entries = poifsDir.getEntries(); entries.hasNext(); ) {
      Entry entry = (Entry)entries.next();
      
      if(entry.getName().equals("Workbook")) {
        return new ExcelExtractor(poifsDir, fs);
      }
      if(entry.getName().equals("WordDocument")) {
        return new WordExtractor(poifsDir, fs);
      }
      if(entry.getName().equals("PowerPoint Document")) {

View Full Code Here

Examples of org.apache.poi.hssf.extractor.ExcelExtractor

          assertEquals("Top Right Cell", s.getRow(0).getCell(255).getStringCellValue());
          assertEquals("Bottom Left Cell", s.getRow(65535).getCell(0).getStringCellValue());
          assertEquals("Bottom Right Cell", s.getRow(65535).getCell(255).getStringCellValue());
          
          // Extract and check
          ExcelExtractor ex = new ExcelExtractor(wb);
          String text = ex.getText();
          assertTrue(text.contains("Top Left Cell"));
          assertTrue(text.contains("Top Right Cell"));
          assertTrue(text.contains("Bottom Left Cell"));
          assertTrue(text.contains("Bottom Right Cell"));
       }

View Full Code Here

Examples of org.apache.poi.hssf.extractor.ExcelExtractor

  
  public IndexDocument getIndexedDocument(File2Index fileData)
      throws SolrException {
    try {
      POIFSFileSystem fs = new POIFSFileSystem(new ByteArrayInputStream(fileData.data));
      ExcelExtractor extractor = new ExcelExtractor(fs);
      String excelText = extractor.getText();


      return new IndexDocument(fileData.path, excelText, null);
    } catch (IOException e) {
      String msg = "Failed to write to the index";
      log.error(msg, e);

View Full Code Here

Examples of org.apache.poi.hssf.extractor.ExcelExtractor

                resourceURL = new URL(url);
                is = resourceURL.openStream();
            }


            POIFSFileSystem fs = new POIFSFileSystem(is);
            ExcelExtractor extractor = new ExcelExtractor(fs);
            String excelText = extractor.getText();


            Document document = new Document();
            document.add(new Field("id", id, Field.Store.YES, Field.Index.TOKENIZED));
            document.add(
                    new Field("content", excelText, Field.Store.NO, Field.Index.TOKENIZED));

View Full Code Here

Examples of org.apache.poi.hssf.extractor.ExcelExtractor

          assertEquals("Top Right Cell", s.getRow(0).getCell(255).getStringCellValue());
          assertEquals("Bottom Left Cell", s.getRow(65535).getCell(0).getStringCellValue());
          assertEquals("Bottom Right Cell", s.getRow(65535).getCell(255).getStringCellValue());
          
          // Extract and check
          ExcelExtractor ex = new ExcelExtractor(wb);
          String text = ex.getText();
          assertTrue(text.contains("Top Left Cell"));
          assertTrue(text.contains("Top Right Cell"));
          assertTrue(text.contains("Bottom Left Cell"));
          assertTrue(text.contains("Bottom Right Cell"));
       }

View Full Code Here

Examples of org.apache.poi.hssf.extractor.ExcelExtractor

          assertEquals("Top Right Cell", s.getRow(0).getCell(255).getStringCellValue());
          assertEquals("Bottom Left Cell", s.getRow(65535).getCell(0).getStringCellValue());
          assertEquals("Bottom Right Cell", s.getRow(65535).getCell(255).getStringCellValue());
          
          // Extract and check
          ExcelExtractor ex = new ExcelExtractor(wb);
          String text = ex.getText();
          assertTrue(text.contains("Top Left Cell"));
          assertTrue(text.contains("Top Right Cell"));
          assertTrue(text.contains("Bottom Left Cell"));
          assertTrue(text.contains("Bottom Right Cell"));
          ex.close();
       }
    }

View Full Code Here

Examples of org.apache.poi.hssf.extractor.ExcelExtractor

   */
  public void testComparedToOLE2() throws Exception {
    // A fairly simple file - ooxml
     XSSFEventBasedExcelExtractor ooxmlExtractor = getExtractor("SampleSS.xlsx");


    ExcelExtractor ole2Extractor =
      new ExcelExtractor(HSSFTestDataSamples.openSampleWorkbook("SampleSS.xls"));
    
    POITextExtractor[] extractors =
      new POITextExtractor[] { ooxmlExtractor, ole2Extractor };
    for (int i = 0; i < extractors.length; i++) {
      POITextExtractor extractor = extractors[i];
      
      String text = extractor.getText().replaceAll("[\r\t]", "");
      assertTrue(text.startsWith("First Sheet\nTest spreadsheet\n2nd row2nd row 2nd column\n"));
      Pattern pattern = Pattern.compile(".*13(\\.0+)?\\s+Sheet3.*", Pattern.DOTALL);
      Matcher m = pattern.matcher(text);
      assertTrue(m.matches());      
    }
    
    ole2Extractor.close();
    ooxmlExtractor.close();
  }

View Full Code Here

Examples of org.apache.poi.hssf.extractor.ExcelExtractor

   */
  public void testComparedToOLE2() throws IOException {
    // A fairly simple file - ooxml
    XSSFExcelExtractor ooxmlExtractor = getExtractor("SampleSS.xlsx");


    ExcelExtractor ole2Extractor =
      new ExcelExtractor(HSSFTestDataSamples.openSampleWorkbook("SampleSS.xls"));
    
    POITextExtractor[] extractors =
      new POITextExtractor[] { ooxmlExtractor, ole2Extractor };
    for (int i = 0; i < extractors.length; i++) {
      POITextExtractor extractor = extractors[i];
      
      String text = extractor.getText().replaceAll("[\r\t]", "");
      assertTrue(text.startsWith("First Sheet\nTest spreadsheet\n2nd row2nd row 2nd column\n"));
      Pattern pattern = Pattern.compile(".*13(\\.0+)?\\s+Sheet3.*", Pattern.DOTALL);
      Matcher m = pattern.matcher(text);
      assertTrue(m.matches());      
    }


    ole2Extractor.close();
    ooxmlExtractor.close();
  }

View Full Code Here

Examples of org.apache.poi.hssf.extractor.ExcelExtractor

          assertEquals("Top Right Cell", s.getRow(0).getCell(255).getStringCellValue());
          assertEquals("Bottom Left Cell", s.getRow(65535).getCell(0).getStringCellValue());
          assertEquals("Bottom Right Cell", s.getRow(65535).getCell(255).getStringCellValue());
          
          // Extract and check
          ExcelExtractor ex = new ExcelExtractor(wb);
          String text = ex.getText();
          assertTrue(text.contains("Top Left Cell"));
          assertTrue(text.contains("Top Right Cell"));
          assertTrue(text.contains("Bottom Left Cell"));
          assertTrue(text.contains("Bottom Right Cell"));
          ex.close();
       }
    }

View Full Code Here

0 1 2 3

TOP

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.