The event API is brand new. It is intended for intermediate developers who are willing to learn a little bit of the low level API structures. Its relatively simple to use, but requires a basic understanding of the parts of an Excel file (or willingness to learn). The advantage provided is that you can read an XLS with a relatively small memory footprint.

To use this API you construct an instance of org.apache.poi.hssf.eventmodel.HSSFRequest. Register a class you create that supports the org.apache.poi.hssf.eventmodel.HSSFListener interface using the HSSFRequest.addListener(yourlistener, recordsid). The recordsid should be a static reference number (such as BOFRecord.sid) contained in the classes in org.apache.poi.hssf.record. The trick is you have to know what these records are. Alternatively you can call HSSFRequest.addListenerForAllRecords(mylistener). In order to learn about these records you can either read all of the javadoc in the org.apache.poi.hssf.record package or you can just hack up a copy of org.apache.poi.hssf.dev.EFHSSF and adapt it to your needs. TODO: better documentation on records.

Once you've registered your listeners in the HSSFRequest object you can construct an instance of org.apache.poi.poifs.filesystem.FileSystem (see POIFS howto) and pass it your XLS file inputstream. You can either pass this, along with the request you constructed, to an instance of HSSFEventFactory via the HSSFEventFactory.processWorkbookEvents(request, Filesystem) method, or you can get an instance of DocumentInputStream from Filesystem.createDocumentInputStream("Workbook") and pass it to HSSFEventFactory.processEvents(request, inputStream). Once you make this call, the listeners that you constructed receive calls to their processRecord(Record) methods with each Record they are registered to listen for until the file has been completely read.

A code excerpt from org.apache.poi.hssf.dev.EFHSSF (which is in CVS or the source distribution) is reprinted below with excessive comments:

/**
 * This example shows how to use the event API for reading a file.
 */
public class EventExample
        implements HSSFListener
{
    private SSTRecord sstrec;

    /**
     * This method listens for incoming records and handles them as required.
     * @param record    The record that was found while reading.
     */
    public void processRecord(Record record)
    {
        switch (record.getSid())
        {
            // the BOFRecord can represent either the beginning of a sheet or the workbook
            case BOFRecord.sid:
                BOFRecord bof = (BOFRecord) record;
                if (bof.getType() == bof.TYPE_WORKBOOK)
                {
                    System.out.println("Encountered workbook");
                    // assigned to the class level member
                } else if (bof.getType() == bof.TYPE_WORKSHEET)
                {
                    System.out.println("Encountered sheet reference");
                }
                break;
            case BoundSheetRecord.sid:
                BoundSheetRecord bsr = (BoundSheetRecord) record;
                System.out.println("New sheet named: " + bsr.getSheetname());
                break;
            case RowRecord.sid:
                RowRecord rowrec = (RowRecord) record;
                System.out.println("Row found, first column at "
                        + rowrec.getFirstCol() + " last column at " + rowrec.getLastCol());
                break;
            case NumberRecord.sid:
                NumberRecord numrec = (NumberRecord) record;
                System.out.println("Cell found with value " + numrec.getValue()
                        + " at row " + numrec.getRow() + " and column " + numrec.getColumn());
                break;
                // SSTRecords store a array of unique strings used in Excel.
            case SSTRecord.sid:
                sstrec = (SSTRecord) record;
                for (int k = 0; k < sstrec.getNumUniqueStrings(); k++)
                {
                    System.out.println("String table value " + k + " = " + sstrec.getString(k));
                }
                break;
            case LabelSSTRecord.sid:
                LabelSSTRecord lrec = (LabelSSTRecord) record;
                System.out.println("String cell found with value "
                        + sstrec.getString(lrec.getSSTIndex()));
                break;
        }
    }

    /**
     * Read an excel file and spit out what we find.
     *
     * @param args      Expect one argument that is the file to read.
     * @throws IOException  When there is an error processing the file.
     */
    public static void main(String[] args) throws IOException
    {
        // create a new file input stream with the input file specified
        // at the command line
        FileInputStream fin = new FileInputStream(args[0]);
        // create a new org.apache.poi.poifs.filesystem.Filesystem
        POIFSFileSystem poifs = new POIFSFileSystem(fin);
        // get the Workbook (excel part) stream in a InputStream
        InputStream din = poifs.createDocumentInputStream("Workbook");
        // construct out HSSFRequest object
        HSSFRequest req = new HSSFRequest();
        // lazy listen for ALL records with the listener shown above
        req.addListenerForAllRecords(new EventExample());
        // create our event factory
        HSSFEventFactory factory = new HSSFEventFactory();
        // process our events based on the document input stream
        factory.processEvents(req, din);
        // once all the events are processed close our file input stream
        fin.close();
        // and our document input stream (don't want to leak these!)
        din.close();
        System.out.println("done.");
    }
}

The low level API is not much to look at. It consists of lots of "Records" in the org.apache.poi.hssf.record.* package, and set of helper classes in org.apache.poi.hssf.model.*. The record classes are consistent with the low level binary structures inside a BIFF8 file (which is embedded in a POIFS file system). You probably need the book: "Microsoft Excel 97 Developer's Kit" from Microsoft Press in order to understand how these fit together (out of print but easily obtainable from Amazon's used books). In order to gain a good understanding of how to use the low level APIs should view the source in org.apache.poi.hssf.usermodel.* and the classes in org.apache.poi.hssf.model.*. You should read the documentation for the POIFS libraries as well.

The HSSF application is nothing more than a test for the high level API (and indirectly the low level support). The main body of its code is repeated above. To run it:

• download the poi-alpha build and untar it (tar xvzf tarball.tar.gz)
• set up your classpath as follows: export HSSFDIR={wherever you put HSSF's jar files} export LOG4JDIR={wherever you put LOG4J's jar files} export CLASSPATH=$CLASSPATH:$HSSFDIR/hssf.jar:$HSSFDIR/poi-poifs.jar:$HSSFDIR/poi-util.jar:$LOG4JDIR/jog4j.jar
• type: java org.apache.poi.hssf.dev.HSSF ~/myxls.xls write

This should generate a test sheet in your home directory called "myxls.xls".

• Type: java org.apache.poi.hssf.dev.HSSF ~/input.xls output.xls

  This is the read/write/modify test. It reads in the spreadsheet, modifies a cell, and writes it back out. Failing this test is not necessarily a bad thing. If HSSF tries to modify a non-existant sheet then this will most likely fail. No big deal.

HSSF now has a logging facility (using commons logging) that will record massive amounts of debugging information. Its mostly useful to us hssf-developing geeks, but might be useful in tracking down problems.

So Why use commons logging rather than log4j? Well the following discussion from the jakarta-general mailing list sums it up pretty well. (Thanks Morgan)

Here's the problem, as I see it.

Suppose Commons component A decides to adopt Log4J, Commons component B decides to adopt LogKit, and Commons component C adopts JDK1.4 logging. They will all minimally function with the right jars in the classpath. However you (the end-user) are left with maintaining configuration for 3 different logging APIs, which is tedious at best. When you take into account cool features like variable log levels, Log4J appenders and the like, you're pretty much guaranteed to swallow up useful configuration options because sophisticated configurations are too difficult to maintain over mutiple logging implementations.

Contrarily, if all three Commons components use a logging facade, you can focus all your configuration efforts on one logging implementation. Sure, there is a trade-off; you don't have access to all the features, and the interface between the facade and the implementation must be maintained. But the benefits are not just political; they potentially make the end-users configuration much easier.

Even if all Commons components used the same logging implementation (Log4J for example), other projects in Jakarta-land may choose otherwise. If you add enough Jakarta projects to your environment, you eventually end up with the scenario described above. It's a worthwhile effort to attempt a logging solution that plays well with the Jakarta community at large. I think in many cases the Commons Logging component can fill that role.

Refer to the commons logging package level javadoc for more information concerning how to configure commons logging.

HSSF has a number of tools useful for developers to debug/develop stuff using HSSF (and more generally XLS files). We've already discussed the app for testing HSSF read/write/modify capabilities; now we'll talk a bit about BiffViewer. Early on in the development of HSSF, it was decided that knowing what was in a record, what was wrong with it, etc. was virtually impossible with the available tools. So we developed BiffViewer. You can find it at org.apache.poi.hssf.dev.BiffViewer. It performs two basic functions and a derivative.

The first is "biffview". To do this you run it (assumes you have everything setup in your classpath and that you know what you're doing enough to be thinking about this) with an xls file as a parameter. It will give you a listing of all understood records with their data and a list of not-yet-understood records with no data (because it doesn't know how to interpret them). This listing is useful for several things. First, you can look at the values and SEE what is wrong in quasi-English. Second, you can send the output to a file and compare it.

The second function is "big freakin dump", just pass a file and a second argument matching "bfd" exactly. This will just make a big hexdump of the file.

Lastly, there is "mixed" mode which does the same as regular biffview, only it includes hex dumps of certain records intertwined. To use that just pass a file with a second argument matching "on" exactly.

In the next release cycle we'll also have something called a FormulaViewer. The class is already there, but its not very useful yet. When it does something, we'll document it.

This release contains code that supports "internationalization" or more accurately non-US/UK languages; however, it has not been tested with the new API changes (please help us with this). We've shifted focus a bit for this release in recognition of the international support we've gotten. We're going to focus on western European languages for our first beta. We're more than happy to accept help in supporting non-Western European languages if someone who knows what they're doing in this area is willing to pitch in! (There is next to no documentation on what is necessary to support such a move and its really hard to support a language when you don't even know the alphabet).

This release of HSSF does not yet support Formulas. I've been focusing on the requests I've gotten in. That being said, if we get more user feedback on what is most useful first we'll aim for that. As a general principal, HSSF's goal is to support HSSF-Serializer (meaning an emphasis on write). We would like to hear from you! How are you using HSSF/POIFS? How would you like to use it? What features are most important first?