Shinsuke Sugaya <******@****.***> 1444529815 +0900

Shinsuke Sugaya <******@****.***> 1444529815 +0900

    }

    public void test_getText_msexcelx() {
        InputStream in = ResourceUtil.getResourceAsStream("extractor/msoffice/test.xlsx");
        Map<String, String> params = new HashMap<String, String>();
        params.put("resourceName", "test.xlsx");
        ExtractData extractData = jodExtractor.getText(in, params);
        String content = extractData.getContent();
        CloseableUtil.closeQuietly(in);

        // Spreadsheet Formats
        extensionMap.put("ods", "tsv");
        extensionMap.put("ots", "tsv");
        extensionMap.put("sxc", "tsv");
        extensionMap.put("xls", "tsv");
        extensionMap.put("xlsx", "tsv");
        extensionMap.put("csv", "tsv");
        extensionMap.put("tsv", "tsv");
        // Presentation Formats
        extensionMap.put("odp", "pdf");
        extensionMap.put("otp", "pdf");

/- - - - - - - - - - - - - - - - - - - -
playsql
  |-data
     |-common
     |  |-xls
     |     |-10-master.xls
     |     |-defaultValueMap.dataprop
     |-ut
        |-xls
           |-20-member.xls  
           |-30-product.xls  
           |-defaultValueMap.dataprop
- - - - - - - - - -/

The format of a xls file is like this:
/- - - - - - - - - - - - - - - - - - - -
|MEMBER_ID|MEMBER_NAME|BIRTHDATE |

        msExcelExtractor = container.getComponent("msExcelExtractor");
    }

    public void test_getText() {
        final InputStream in = ResourceUtil.getResourceAsStream("extractor/msoffice/test.xls");
        final String content = msExcelExtractor.getText(in, null).getContent();
        CloseableUtil.closeQuietly(in);
        logger.info(content);
        assertTrue(content.contains("テスト"));
    }

application/access             mdf              # Microsoft Access
#application/excel              xls              # Microsoft Excel
application/vnd.ms-excel       xls              # Microsoft Excel
application/font-tdpfr         pfr              # TrueDoc Portable Font Resource
application/futuresplash       spl              # Macromedia Flash
application/hep                hep              # Hummingbird Host Explorer Profiles

import org.codelibs.fess.crawler.entity.ExtractData;
import org.codelibs.fess.crawler.exception.CrawlerSystemException;
import org.codelibs.fess.crawler.exception.ExtractException;

/**
 * Gets a text from .xls file.
 *
 * @author shinsuke
 *
 */
public class MsExcelExtractor extends AbstractExtractor {
    /*
     * (non-Javadoc)
     *
     * @see

    // public ResponseData call() throws Exception {
    // String[] urls =
    // new String[] {
    // "http://.../",
    // "http://.../test.pdf",
    // "http://.../test.doc",
    // "http://.../test.xls",
    // "http://.../test.ppt",
    // "http://.../test.txt", };
    // for (String url : urls) {
    // ResponseData responseData = httpClient.doGet(url);
    // // assertEquals(200, responseData.getHttpStatusCode());