<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<html>
<head>
<META http-equiv="Content-Type" content="text/html; charset=UTF-8">
<!--*** This is a generated file. Do not edit. ***-->
<link rel="stylesheet" href="../skin/tigris.css" type="text/css">
<link rel="stylesheet" href="../skin/mysite.css" type="text/css">
<link rel="stylesheet" href="../skin/site.css" type="text/css">
<link media="print" rel="stylesheet" href="../skin/print.css" type="text/css">
<title>HPSF HOW-TO</title>
</head>
<body bgcolor="white" class="composite">
<!--================= start Banner ==================-->
<div id="banner">
<table width="100%" cellpadding="8" cellspacing="0" summary="banner" border="0">
<tbody>
<tr>
<!--================= start Group Logo ==================-->
<td width="50%" align="left">
<div class="groupLogo">
<a href="http://poi.apache.org"><img border="0" class="logoImage" alt="Apache POI" src="../resources/images/group-logo.jpg"></a>
</div>
</td>
<!--================= end Group Logo ==================-->
<!--================= start Project Logo ==================--><td width="50%" align="right">
<div align="right" class="projectLogo">
<a href="http://poi.apache.org/"><img border="0" class="logoImage" alt="POI" src="../resources/images/project-logo.jpg"></a>
</div>
</td>
<!--================= end Project Logo ==================-->
</tr>
</tbody>
</table>
</div>
<!--================= end Banner ==================-->
<!--================= start Main ==================-->
<table width="100%" cellpadding="0" cellspacing="0" border="0" summary="nav" id="breadcrumbs">
<tbody>
<!--================= start Status ==================-->
<tr class="status">
<td>
<!--================= start BreadCrumb ==================--><a href="http://www.apache.org/">Apache</a> | <a href="http://poi.apache.org/">POI</a><a href=""></a>
<!--================= end BreadCrumb ==================--></td><td id="tabs">
<!--================= start Tabs ==================-->
<div class="tab">
<span class="selectedTab"><a class="base-selected" href="../index.html">Home</a></span> | <script language="Javascript" type="text/javascript">
function printit() {
if (window.print) {
window.print() ;
} else {
var WebBrowser = '<OBJECT ID="WebBrowser1" WIDTH="0" HEIGHT="0" CLASSID="CLSID:8856F961-340A-11D0-A96B-00C04FD705A2"></OBJECT>';
document.body.insertAdjacentHTML('beforeEnd', WebBrowser);
WebBrowser1.ExecWB(6, 2);//Use a 1 vs. a 2 for a prompting dialog box WebBrowser1.outerHTML = "";
}
}
</script><script language="Javascript" type="text/javascript">
var NS = (navigator.appName == "Netscape");
var VERSION = parseInt(navigator.appVersion);
if (VERSION > 3) {
document.write(' <a title="PRINT this page OUT" href="javascript:printit()">PRINT</a>');
}
</script>
</div>
<!--================= end Tabs ==================-->
</td>
</tr>
</tbody>
</table>
<!--================= end Status ==================-->
<table id="main" width="100%" cellpadding="8" cellspacing="0" summary="" border="0">
<tbody>
<tr valign="top">
<!--================= start Menu ==================-->
<td id="leftcol">
<div id="navcolumn">
<div class="menuBar">
<div class="menu">
<span class="menuLabel">Apache POI</span>
<div class="menuItem">
<a href="../index.html">Top</a>
</div>
</div>
<div class="menu">
<span class="menuLabel">HPSF</span>
<div class="menuItem">
<a href="index.html">Overview</a>
</div>
<div class="menuItem">
<span class="menuSelected">How To</span>
</div>
<div class="menuItem">
<a href="thumbnails.html">Thumbnails</a>
</div>
<div class="menuItem">
<a href="internals.html">Internals</a>
</div>
<div class="menuItem">
<a href="todo.html">To Do</a>
</div>
</div>
</div>
</div>
<form target="_blank" action="http://www.google.com/search" method="get">
<table summary="search" border="0" cellspacing="0" cellpadding="0">
<tr>
<td><img height="1" width="1" alt="" src="../skin/images/spacer.gif" class="spacer"></td><td nowrap="nowrap">
Search Apache POI<br>
<input value="poi.apache.org" name="sitesearch" type="hidden"><input size="10" name="q" id="query" type="text"><img height="1" width="5" alt="" src="../skin/images/spacer.gif" class="spacer"><input name="Search" value="GO" type="submit"></td><td><img height="1" width="1" alt="" src="../skin/images/spacer.gif" class="spacer"></td>
</tr>
<tr>
<td colspan="3"><img height="7" width="1" alt="" src="../skin/images/spacer.gif" class="spacer"></td>
</tr>
<tr>
<td class="bottom-left-thick"></td><td bgcolor="#a5b6c6"><img height="1" width="1" alt="" src="../skin/images/spacer.gif" class="spacer"></td><td class="bottom-right-thick"></td>
</tr>
</table>
</form>
</td>
<!--================= end Menu ==================-->
<!--================= start Content ==================--><td>
<div id="bodycol">
<div class="app">
<div align="center">
<h1>HPSF HOW-TO</h1>
</div>
<div class="h3">
<a name="How+To+Use+the+HPSF+API"></a>
<div class="h3">
<h3>How To Use the HPSF API</h3>
</div>
<p>This HOW-TO is organized in four sections. You should read them
sequentially because the later sections build upon the earlier ones.</p>
<ol>
<li>
The <a href="#sec1">first section</a> explains how to <strong>read
the most important standard properties</strong> of a Microsoft Office
document. Standard properties are things like title, author, creation
date etc. It is quite likely that you will find here what you need and
don't have to read the other sections.
</li>
<li>
The <a href="#sec2">second section</a> goes a small step
further and focusses on <strong>reading additional standard
properties</strong>. It also talks about <strong>exceptions</strong> that
may be thrown when dealing with HPSF and shows how you can <strong>read
properties of embedded objects</strong>.
</li>
<li>
The <a href="#sec3">third section</a> explains how to <strong>write
standard properties</strong>. HPSF provides some high-level classes and
methods which make writing of standard properties easy. They are based on
the low-level writing functions explained in the <a href="#sec3">fifth
section</a>.
</li>
<li>
The <a href="#sec4">fourth section</a> tells how to <strong>read
non-standard properties</strong>. Non-standard properties are
application-specific triples consisting of an ID, a type, and a value.
</li>
<li>
The <a href="#sec5">fifth section</a> tells you how to <strong>write
property set streams</strong> using HPSF's low-level methods. You have to
understand the <a href="#sec3">fourth section</a> before you should
think about low-level writing properties. Check the Javadoc API
documentation to find out about the details!
</li>
</ol>
<div class="frame note">
<div class="label">Note</div>
<div class="content">
<strong>Please note:</strong> HPSF's writing functionality is
<strong>not</strong> present in POI releases up to and including 2.5. In
order to write properties you have to download a 3.0.x POI release,
or retrieve the POI development version from the <a href="../subversion.html">Subversion repository</a>.</div>
</div>
<a name="sec1"></a>
<a name="Reading+Standard+Properties"></a>
<div class="h4">
<h4>Reading Standard Properties</h4>
</div>
<div class="frame note">
<div class="label">Note</div>
<div class="content">This section explains how to read the most important standard
properties of a Microsoft Office document. Standard properties are things
like title, author, creation date etc. This section introduces the
<strong>summary information stream</strong> which is used to keep these
properties. Chances are that you will find here what you need and don't
have to read the other sections.</div>
</div>
<p>If all you are interested in is getting the textual content of
all the document properties, such as for full text indexing, then
take a look at
<span class="codefrag">org.apache.poi.hpsf.extractor.HPSFPropertiesExtractor</span>. However,
if you want full access to the properties, please read on!</p>
<p>The first thing you should understand is that a Microsoft Office file is
not one large bunch of bytes but has an internal filesystem structure with
files and directories. You can access these files and directories using
the <a href="../poifs/index.html">POI filesystem (POIFS)</a>
provides. A file or document in a POI filesystem is also called a
<strong>stream</strong> - The properties of, say, an Excel document are
stored apart of the actual spreadsheet data in separate streams. The good
new is that this separation makes the properties independent of the
concrete Microsoft Office file. In the following text we will always say
"POI filesystem" instead of "Microsoft Office file" because a POI
filesystem is not necessarily created by or for a Microsoft Office
application, because it is shorter, and because we want to avoid the name
of That Redmond Company.</p>
<p>The following example shows how to read the "title" property. Reading
other properties is similar. Consider the API documentation of the class
<span class="codefrag">org.apache.poi.hpsf.SummaryInformation</span> to learn which methods
are available.</p>
<p>The standard properties this section focusses on can be found in a
document called <em>\005SummaryInformation</em> located in the root of the
POI filesystem. The notation <em>\005</em> in the document's name means
the character with a decimal value of 5. In order to read the "title"
property, an application has to perform the following steps:</p>
<ol>
<li>
Open the document <em>\005SummaryInformation</em> located in the root
of the POI filesystem.
</li>
<li>
Create an instance of the class <span class="codefrag">SummaryInformation</span> from
that document.
</li>
<li>
Call the <span class="codefrag">SummaryInformation</span> instance's
<span class="codefrag">getTitle()</span> method.
</li>
</ol>
<p>Sounds easy, doesn't it? Here are the steps in detail.</p>
<a name="Open+the+document+%5C005SummaryInformation+in+the+root+of+the%0A+++++++POI+filesystem"></a>
<div class="h2">
<h2>Open the document \005SummaryInformation in the root of the
POI filesystem</h2>
</div>
<p>An application that wants to open a document in a POI filesystem
(POIFS) proceeds as shown by the following code fragment. The full
source code of the sample application is available in the
<em>examples</em> section of the POI source tree as
<em>ReadTitle.java</em>.</p>
<pre class="code">
import java.io.*;
import org.apache.poi.hpsf.*;
import org.apache.poi.poifs.eventfilesystem.*;
// ...
public static void main(String[] args)
throws IOException
{
final String filename = args[0];
POIFSReader r = new POIFSReader();
r.registerListener(new MyPOIFSReaderListener(),
"\005SummaryInformation");
r.read(new FileInputStream(filename));
}</pre>
<p>The first interesting statement is</p>
<pre class="code">POIFSReader r = new POIFSReader();</pre>
<p>It creates a
<span class="codefrag">org.apache.poi.poifs.eventfilesystem.POIFSReader</span> instance
which we shall need to read the POI filesystem. Before the application
actually opens the POI filesystem we have to tell the
<span class="codefrag">POIFSReader</span> which documents we are interested in. In this
case the application should do something with the document
<em>\005SummaryInformation</em>.</p>
<pre class="code">
r.registerListener(new MyPOIFSReaderListener(),
"\005SummaryInformation");</pre>
<p>This method call registers a
<span class="codefrag">org.apache.poi.poifs.eventfilesystem.POIFSReaderListener</span>
with the <span class="codefrag">POIFSReader</span>. The <span class="codefrag">POIFSReaderListener</span>
interface specifies the method <span class="codefrag">processPOIFSReaderEvent()</span>
which processes a document. The class
<span class="codefrag">MyPOIFSReaderListener</span> implements the
<span class="codefrag">POIFSReaderListener</span> and thus the
<span class="codefrag">processPOIFSReaderEvent()</span> method. The eventing POI
filesystem calls this method when it finds the
<em>\005SummaryInformation</em> document. In the sample application
<span class="codefrag">MyPOIFSReaderListener</span> is a static class in the
<em>ReadTitle.java</em> source file.</p>
<p>Now everything is prepared and reading the POI filesystem can
start:</p>
<pre class="code">r.read(new FileInputStream(filename));</pre>
<p>The following source code fragment shows the
<span class="codefrag">MyPOIFSReaderListener</span> class and how it retrieves the
title.</p>
<pre class="code">
static class MyPOIFSReaderListener implements POIFSReaderListener
{
public void processPOIFSReaderEvent(POIFSReaderEvent event)
{
SummaryInformation si = null;
try
{
si = (SummaryInformation)
PropertySetFactory.create(event.getStream());
}
catch (Exception ex)
{
throw new RuntimeException
("Property set stream \"" +
event.getPath() + event.getName() + "\": " + ex);
}
final String title = si.getTitle();
if (title != null)
System.out.println("Title: \"" + title + "\"");
else
System.out.println("Document has no title.");
}
}
</pre>
<p>The line</p>
<pre class="code">SummaryInformation si = null;</pre>
<p>declares a <span class="codefrag">SummaryInformation</span> variable and initializes it
with <span class="codefrag">null</span>. We need an instance of this class to access the
title. The instance is created in a <span class="codefrag">try</span> block:</p>
<pre class="code">si = (SummaryInformation)
PropertySetFactory.create(event.getStream());</pre>
<p>The expression <span class="codefrag">event.getStream()</span> returns the input stream
containing the bytes of the property set stream named
<em>\005SummaryInformation</em>. This stream is passed into the
<span class="codefrag">create</span> method of the factory class
<span class="codefrag">org.apache.poi.hpsf.PropertySetFactory</span> which returns
a <span class="codefrag">org.apache.poi.hpsf.PropertySet</span> instance. It is more or
less safe to cast this result to <span class="codefrag">SummaryInformation</span>, a
convenience class with methods like <span class="codefrag">getTitle()</span>,
<span class="codefrag">getAuthor()</span> etc.</p>
<p>The <span class="codefrag">PropertySetFactory.create()</span> method may throw all sorts
of exceptions. We'll deal with them in the next sections. For now we just
catch all exceptions and throw a <span class="codefrag">RuntimeException</span>
containing the message text of the origin exception.</p>
<p>If all goes well, the sample application retrieves the title and prints
it to the standard output. As you can see you must be prepared for the
case that the POI filesystem does not have a title.</p>
<pre class="code">final String title = si.getTitle();
if (title != null)
System.out.println("Title: \"" + title + "\"");
else
System.out.println("Document has no title.");</pre>
<p>Please note that a POI filesystem does not necessarily contain the
<em>\005SummaryInformation</em> stream. The documents created by the
Microsoft Office suite have one, as far as I know. However, an Excel
spreadsheet exported from StarOffice 5.2 won't have a
<em>\005SummaryInformation</em> stream. In this case the applications
won't throw an exception but simply does not call the
<span class="codefrag">processPOIFSReaderEvent</span> method. You have been warned!</p>
<a name="sec2"></a>
<a name="Additional+Standard+Properties%2C+Exceptions+And+Embedded%0A++++Objects"></a>
<div class="h4">
<h4>Additional Standard Properties, Exceptions And Embedded
Objects</h4>
</div>
<div class="frame note">
<div class="label">Note</div>
<div class="content">This section focusses on reading additional standard properties which
are kept in the <strong>document summary information</strong> stream. It
also talks about exceptions that may be thrown when dealing with HPSF and
shows how you can read properties of embedded objects.</div>
</div>
<p>A couple of <strong>additional standard properties</strong> are not
contained in the <em>\005SummaryInformation</em> stream explained
above. Examples for such properties are a document's category or the
number of multimedia clips in a PowerPoint presentation. Microsoft has
invented an additional stream named
<em>\005DocumentSummaryInformation</em> to hold these properties. With two
minor exceptions you can proceed exactly as described above to read the
properties stored in <em>\005DocumentSummaryInformation</em>:</p>
<ul>
<li>Instead of <em>\005SummaryInformation</em> use
<em>\005DocumentSummaryInformation</em> as the stream's name.</li>
<li>Replace all occurrences of the class
<span class="codefrag">SummaryInformation</span> by
<span class="codefrag">DocumentSummaryInformation</span>.</li>
</ul>
<p>And of course you cannot call <span class="codefrag">getTitle()</span> because
<span class="codefrag">DocumentSummaryInformation</span> has different query methods,
e.g. <span class="codefrag">getCategory</span>. See the Javadoc API documentation for the
details.</p>
<p>In the previous section the application simply caught all
<strong>exceptions</strong> and was in no way interested in any
details. However, a real application will likely want to know what went
wrong and act appropriately. Besides any I/O exceptions there are three
HPSF resp. POI specific exceptions you should know about:</p>
<dl>
<dt>
<span class="codefrag">NoPropertySetStreamException</span>:</dt>
<dd>
This exception is thrown if the application tries to create a
<span class="codefrag">PropertySet</span> instance from a stream that is not a
property set stream. (<span class="codefrag">SummaryInformation</span> and
<span class="codefrag">DocumentSummaryInformation</span> are subclasses of
<span class="codefrag">PropertySet</span>.) A faulty property set stream counts as not
being a property set stream at all. An application should be prepared to
deal with this case even if it opens streams named
<em>\005SummaryInformation</em> or
<em>\005DocumentSummaryInformation</em>. These are just names. A
stream's name by itself does not ensure that the stream contains the
expected contents and that this contents is correct.
</dd>
<dt>
<span class="codefrag">UnexpectedPropertySetTypeException</span>
</dt>
<dd>This exception is thrown if a certain type of property set is
expected somewhere (e.g. a <span class="codefrag">SummaryInformation</span> or
<span class="codefrag">DocumentSummaryInformation</span>) but the provided property
set is not of that type.</dd>
<dt>
<span class="codefrag">MarkUnsupportedException</span>
</dt>
<dd>This exception is thrown if an input stream that is to be parsed
into a property set does not support the
<span class="codefrag">InputStream.mark(int)</span> operation. The POI filesystem uses
the <span class="codefrag">DocumentInputStream</span> class which does support this
operation, so you are safe here. However, if you read a property set
stream from another kind of input stream things may be
different.</dd>
</dl>
<p>Many Microsoft Office documents contain <strong>embedded
objects</strong>, for example an Excel sheet within a Word
document. Embedded objects may have property sets of their own. An
application can open these property set streams as described above. The
only difference is that they are not located in the POI filesystem's root
but in a <strong>nested directory</strong> instead. Just register a
<span class="codefrag">POIFSReaderListener</span> for the property set streams you are
interested in. For example, the <em>POIBrowser</em> application
tries to open each and every document in a POI filesystem
as a property set stream. If this operation was successful it displays the
properties.</p>
<a name="sec3"></a>
<a name="Writing+Standard+Properties"></a>
<div class="h4">
<h4>Writing Standard Properties</h4>
</div>
<div class="frame note">
<div class="label">Note</div>
<div class="content">This section explains how to <strong>write standard
properties</strong>. HPSF provides some high-level classes and methods
which make writing of standard properties easy. They are based on the
low-level writing functions explained in <a href="#sec4">another
section</a>.</div>
</div>
<p>As explained above, standard properties are located in the summary
information and document summary information streams of typical POI
filesystems. You have already learned about the classes
<span class="codefrag">SummaryInformation</span> and
<span class="codefrag">DocumentSummaryInformation</span> and their <span class="codefrag">get...()</span>
methods for reading standard properties. These classes also provide
<span class="codefrag">set...()</span> methods for writing properties.</p>
<p>After setting properties in <span class="codefrag">SummaryInformation</span> or
<span class="codefrag">DocumentSummaryInformation</span> you have to write them to a disk
file. The following sample program shows how you can</p>
<ol>
<li>read a disk file into a POI filesystem,</li>
<li>read the document summary information from the POI filesystem,</li>
<li>set a property to a new value,</li>
<li>write the modified document summary information back to the POI
filesystem, and</li>
<li>write the POI filesystem to a disk file.</li>
</ol>
<p>The complete source code of this program is available as
<em>ModifyDocumentSummaryInformation.java</em> in the <em>examples</em>
section of the POI source tree.</p>
<div class="frame note">
<div class="label">Note</div>
<div class="content">Dealing with the summary information stream is analogous to handling
the document summary information and therefore does not need to be
explained here in detailed. See the HPSF API documentation to learn about
the <span class="codefrag">set...()</span> methods of the class
<span class="codefrag">SummaryInformation</span>.</div>
</div>
<p>The first step is to read the POI filesystem into memory:</p>
<pre class="code">InputStream is = new FileInputStream(poiFilesystem);
POIFSFileSystem poifs = new POIFSFileSystem(is);
is.close();</pre>
<p>The code snippet above assumes that the variable
<span class="codefrag">poiFilesystem</span> holds the name of a disk file. It reads the
file from an input stream and creates a <span class="codefrag">POIFSFileSystem</span>
object in memory. After having read the file, the input stream should be
closed as shown.</p>
<p>In order to read the document summary information stream the application
must open the element <em>\005DocumentSummaryInformation</em> in the POI
filesystem's root directory. However, the POI filesystem does not
necessarily contain a document summary information stream, and the
application should be able to deal with that situation. The following
code does so by creating a new <span class="codefrag">DocumentSummaryInformation</span> if
there is none in the POI filesystem:</p>
<pre class="code">DirectoryEntry dir = poifs.getRoot();
DocumentSummaryInformation dsi;
try
{
DocumentEntry dsiEntry = (DocumentEntry)
dir.getEntry(DocumentSummaryInformation.DEFAULT_STREAM_NAME);
DocumentInputStream dis = new DocumentInputStream(dsiEntry);
PropertySet ps = new PropertySet(dis);
dis.close();
dsi = new DocumentSummaryInformation(ps);
}
catch (FileNotFoundException ex)
{
/* There is no document summary information. We have to create a
* new one. */
dsi = PropertySetFactory.newDocumentSummaryInformation();
}
</pre>
<p>In the source code above the statement</p>
<pre class="code">DirectoryEntry dir = poifs.getRoot();</pre>
<p>gets hold of the POI filesystem's root directory as a
<span class="codefrag">DirectoryEntry</span>. The <span class="codefrag">getEntry()</span> method of this
class is used to access a file or directory entry in a directory. However,
if the file to be opened does not exist, a
<span class="codefrag">FileNotFoundException</span> will be thrown. Therefore opening the
document summary information entry should be done in a <span class="codefrag">try</span>
block:</p>
<pre class="code"> DocumentEntry dsiEntry = (DocumentEntry)
dir.getEntry(DocumentSummaryInformation.DEFAULT_STREAM_NAME);</pre>
<p>
<span class="codefrag">DocumentSummaryInformation.DEFAULT_STREAM_NAME</span> represents
the string "\005DocumentSummaryInformation", i.e. the standard name of a
document summary information stream. If this stream exists, the
<span class="codefrag">getEntry()</span> method returns a <span class="codefrag">DocumentEntry</span>. To
read the <span class="codefrag">DocumentEntry</span>'s contents, create a
<span class="codefrag">DocumentInputStream</span>:</p>
<pre class="code"> DocumentInputStream dis = new DocumentInputStream(dsiEntry);</pre>
<p>Up to this point we have used POI's <a href="../poifs/index.html">POIFS component</a>. Now HPSF enters the
stage. A property set is created from the input stream's data:</p>
<pre class="code"> PropertySet ps = new PropertySet(dis);
dis.close();
dsi = new DocumentSummaryInformation(ps); </pre>
<p>If the data really constitutes a property set, a
<span class="codefrag">PropertySet</span> object is created. Otherwise a
<span class="codefrag">NoPropertySetStreamException</span> is thrown. After having read the
data from the input stream the latter should be closed.</p>
<p>Since we know - or at least hope - that the stream named
"\005DocumentSummaryInformation" is not just any property set but really
contains the document summary information, we try to create a new
<span class="codefrag">DocumentSummaryInformation</span> from the property set. If the
stream is not document summary information stream the sample application
fails with a <span class="codefrag">UnexpectedPropertySetTypeException</span>.</p>
<p>If the POI document does not contain a document summary information
stream, we can create a new one in the <span class="codefrag">catch</span> clause. The
<span class="codefrag">PropertySetFactory</span>'s method
<span class="codefrag">newDocumentSummaryInformation()</span> establishes a new and empty
<span class="codefrag">DocumentSummaryInformation</span> instance:</p>
<pre class="code"> dsi = PropertySetFactory.newDocumentSummaryInformation();</pre>
<p>Whether we read the document summary information from the POI filesystem
or created it from scratch, in either case we now have a
<span class="codefrag">DocumentSummaryInformation</span> instance we can write to. Writing
is quite simple, as the following line of code shows:</p>
<pre class="code">dsi.setCategory("POI example");</pre>
<p>This statement sets the "category" property to "POI example". Any
former "category" value will be lost. If there hasn't been a "category"
property yet, a new one will be created.</p>
<p>
<span class="codefrag">DocumentSummaryInformation</span> of course has methods to set the
other standard properties, too - look into the API documentation to see
all of them.</p>
<p>Once all properties are set as needed, they should be stored into the
file on disk. The first step is to write the
<span class="codefrag">DocumentSummaryInformation</span> into the POI filesystem:</p>
<pre class="code">dsi.write(dir, DocumentSummaryInformation.DEFAULT_STREAM_NAME);</pre>
<p>The <span class="codefrag">DocumentSummaryInformation</span>'s <span class="codefrag">write()</span>
method takes two parameters: The first is the <span class="codefrag">DirectoryEntry</span>
in the POI filesystem, the second is the name of the stream to create in
the directory. If this stream already exists, it will be overwritten.</p>
<div class="frame note">
<div class="label">Note</div>
<div class="content">If you not only modified the document summary information but also
the summary information you have to write both of them to the POI
filesystem.</div>
</div>
<p>Still the POI filesystem is a data structure in memory only and must be
written to a disk file to make it permanent. The following lines write
back the POI filesystem to the file it was read from before. Please note
that in production-quality code you should never write directly to the
origin file, because in case of an error everything would be lost. Here it
is done this way to keep the example short.</p>
<pre class="code">OutputStream out = new FileOutputStream(poiFilesystem);
poifs.writeFilesystem(out);
out.close();</pre>
<a name="User-Defined+Properties"></a>
<div class="h2">
<h2>User-Defined Properties</h2>
</div>
<p>If you compare the source code excerpts above with the file containing
the full source code, you will notice that I left out some following
lines of code. The are dealing with the special topic of custom
properties.</p>
<pre class="code">DocumentSummaryInformation dsi = ...
...
CustomProperties customProperties = dsi.getCustomProperties();
if (customProperties == null)
customProperties = new CustomProperties();
/* Insert some custom properties into the container. */
customProperties.put("Key 1", "Value 1");
customProperties.put("Schlüssel 2", "Wert 2");
customProperties.put("Sample Number", new Integer(12345));
customProperties.put("Sample Boolean", new Boolean(true));
customProperties.put("Sample Date", new Date());
/* Read a custom property. */
Object value = customProperties.get("Sample Number");
/* Write the custom properties back to the document summary
* information. */
dsi.setCustomProperties(customProperties);</pre>
<p>Custom properties are properties the user can define himself. Using for
example Microsoft Word he can define these extra properties and give
each of them a <strong>name</strong>, a <strong>type</strong> and a
<strong>value</strong>. The custom properties are stored in the document
information summary along with the standard properties.</p>
<p>The source code example shows how to retrieve the custom properties
as a whole from a <span class="codefrag">DocumentSummaryInformation</span> instance using
the <span class="codefrag">getCustomProperties()</span> method. The result is a
<span class="codefrag">CustomProperties</span> instance or <span class="codefrag">null</span> if no
user-defined properties exist.</p>
<p>Since <span class="codefrag">CustomProperties</span> implements the <span class="codefrag">Map</span>
interface you can read and write properties with the usual
<span class="codefrag">Map</span> methods. However, <span class="codefrag">CustomProperties</span> poses
some restrictions on the types of keys and values.</p>
<ul>
<li>The <strong>key</strong> is a string.</li>
<li>The <strong>value</strong> is one of <span class="codefrag">String</span>,
<span class="codefrag">Boolean</span>, <span class="codefrag">Long</span>, <span class="codefrag">Integer</span>,
<span class="codefrag">Short</span>, or <span class="codefrag">java.util.Date</span>.</li>
</ul>
<p>The <span class="codefrag">CustomProperties</span> class has been designed for easy
access using just keys and values. The underlying Microsoft-specific
custom properties data structure is more complicated. However, it does
not provide noteworthy additional benefits. It is possible to have
multiple properties with the same name or properties without a
name at all. When reading custom properties from a document summary
information stream, the <span class="codefrag">CustomProperties</span> class ignores
properties without a name and keeps only the "last" (whatever that means)
of those properties having the same name. You can find out whether a
<span class="codefrag">CustomProperties</span> instance dropped any properties with the
<span class="codefrag">isPure()</span> method.</p>
<p>You can read and write the full spectrum of custom properties with
HPSF's low-level methods. They are explained in the <a href="#sec4">next section</a>.</p>
<a name="sec4"></a>
<a name="Reading+Non-Standard+Properties"></a>
<div class="h4">
<h4>Reading Non-Standard Properties</h4>
</div>
<div class="frame note">
<div class="label">Note</div>
<div class="content">This section tells how to read non-standard properties. Non-standard
properties are application-specific ID/type/value triples.</div>
</div>
<a name="Overview"></a>
<div class="h2">
<h2>Overview</h2>
</div>
<p>Now comes the real hardcode stuff. As mentioned above,
<span class="codefrag">SummaryInformation</span> and
<span class="codefrag">DocumentSummaryInformation</span> are just special cases of the
general concept of a property set. This concept says that a
<strong>property set</strong> consists of properties and that each
<strong>property</strong> is an entity with an <strong>ID</strong>, a
<strong>type</strong>, and a <strong>value</strong>.</p>
<p>Okay, that was still rather easy. However, to make things more
complicated, Microsoft in its infinite wisdom decided that a property set
shalt be broken into one or more <strong>sections</strong>. Each section
holds a bunch of properties. But since that's still not complicated
enough, a section may have an optional <strong>dictionary</strong> that
maps property IDs to <strong>property names</strong> - we'll explain
later what that means.</p>
<p>The procedure to get to the properties is the following:</p>
<ol>
<li>Use the <strong><span class="codefrag">PropertySetFactory</span></strong> class to
create a <span class="codefrag">PropertySet</span> object from a property set stream. If
you don't know whether an input stream is a property set stream, just
try to call <span class="codefrag">PropertySetFactory.create(java.io.InputStream)</span>:
You'll either get a <span class="codefrag">PropertySet</span> instance returned or an
exception is thrown.</li>
<li>Call the <span class="codefrag">PropertySet</span>'s method <span class="codefrag">getSections()</span>
to get the sections contained in the property set. Each section is
an instance of the <span class="codefrag">Section</span> class.</li>
<li>Each section has a format ID. The format ID of the first section in a
property set determines the property set's type. For example, the first
(and only) section of the summary information property set has a format
ID of <span class="codefrag">F29F85E0-4FF9-1068-AB-91-08-00-2B-27-B3-D9</span>. You can
get the format ID with <span class="codefrag">Section.getFormatID()</span>.</li>
<li>The properties contained in a <span class="codefrag">Section</span> can be retrieved
with <span class="codefrag">Section.getProperties()</span>. The result is an array of
<span class="codefrag">Property</span> instances.</li>
<li>A property has a name, a type, and a value. The <span class="codefrag">Property</span>
class has methods to retrieve them.</li>
</ol>
<a name="A+Sample+Application"></a>
<div class="h2">
<h2>A Sample Application</h2>
</div>
<p>Let's have a look at a sample Java application that dumps all property
set streams contained in a POI file system. The full source code of this
program can be found as <em>ReadCustomPropertySets.java</em> in the
<em>examples</em> area of the POI source code tree. Here are the key
sections:</p>
<pre class="code">import java.io.*;
import java.util.*;
import org.apache.poi.hpsf.*;
import org.apache.poi.poifs.eventfilesystem.*;
import org.apache.poi.util.HexDump;</pre>
<p>The most important package the application needs is
<span class="codefrag">org.apache.poi.hpsf.*</span>. This package contains the HPSF
classes. Most classes named below are from the HPSF package. Of course we
also need the POIFS event file system's classes and <span class="codefrag">java.io.*</span>
since we are dealing with POI I/O. From the <span class="codefrag">java.util</span> package
we use the <span class="codefrag">List</span> and <span class="codefrag">Iterator</span> class. The class
<span class="codefrag">org.apache.poi.util.HexDump</span> provides a methods to dump byte
arrays as nicely formatted strings.</p>
<pre class="code">public static void main(String[] args)
throws IOException
{
final String filename = args[0];
POIFSReader r = new POIFSReader();
/* Register a listener for *all* documents. */
r.registerListener(new MyPOIFSReaderListener());
r.read(new FileInputStream(filename));
}</pre>
<p>The <span class="codefrag">POIFSReader</span> is set up in a way that the listener
<span class="codefrag">MyPOIFSReaderListener</span> is called on every file in the POI file
system.</p>
<a name="The+Property+Set"></a>
<div class="h2">
<h2>The Property Set</h2>
</div>
<p>The listener class tries to create a <span class="codefrag">PropertySet</span> from each
stream using the <span class="codefrag">PropertySetFactory.create()</span> method:</p>
<pre class="code">static class MyPOIFSReaderListener implements POIFSReaderListener
{
public void processPOIFSReaderEvent(POIFSReaderEvent event)
{
PropertySet ps = null;
try
{
ps = PropertySetFactory.create(event.getStream());
}
catch (NoPropertySetStreamException ex)
{
out("No property set stream: \"" + event.getPath() +
event.getName() + "\"");
return;
}
catch (Exception ex)
{
throw new RuntimeException
("Property set stream \"" +
event.getPath() + event.getName() + "\": " + ex);
}
/* Print the name of the property set stream: */
out("Property set stream \"" + event.getPath() +
event.getName() + "\":");</pre>
<p>Creating the <span class="codefrag">PropertySet</span> is done in a <span class="codefrag">try</span>
block, because not each stream in the POI file system contains a property
set. If it is some other file, the
<span class="codefrag">PropertySetFactory.create()</span> throws a
<span class="codefrag">NoPropertySetStreamException</span>, which is caught and
logged. Then the program continues with the next stream. However, all
other types of exceptions cause the program to terminate by throwing a
runtime exception. If all went well, we can print the name of the property
set stream.</p>
<a name="The+Sections"></a>
<div class="h2">
<h2>The Sections</h2>
</div>
<p>The next step is to print the number of sections followed by the
sections themselves:</p>
<pre class="code">/* Print the number of sections: */
final long sectionCount = ps.getSectionCount();
out(" No. of sections: " + sectionCount);
/* Print the list of sections: */
List sections = ps.getSections();
int nr = 0;
for (Iterator i = sections.iterator(); i.hasNext();)
{
/* Print a single section: */
Section sec = (Section) i.next();
// See below for the complete loop body.
}</pre>
<p>The <span class="codefrag">PropertySet</span>'s method <span class="codefrag">getSectionCount()</span>
returns the number of sections.</p>
<p>To retrieve the sections, use the <span class="codefrag">getSections()</span>
method. This method returns a <span class="codefrag">java.util.List</span> containing
instances of the <span class="codefrag">Section</span> class in their proper order.</p>
<p>The sample code shows a loop that retrieves the <span class="codefrag">Section</span>
objects one by one and prints some information about each one. Here is
the complete body of the loop:</p>
<pre class="code">/* Print a single section: */
Section sec = (Section) i.next();
out(" Section " + nr++ + ":");
String s = hex(sec.getFormatID().getBytes());
s = s.substring(0, s.length() - 1);
out(" Format ID: " + s);
/* Print the number of properties in this section. */
int propertyCount = sec.getPropertyCount();
out(" No. of properties: " + propertyCount);
/* Print the properties: */
Property[] properties = sec.getProperties();
for (int i2 = 0; i2 < properties.length; i2++)
{
/* Print a single property: */
Property p = properties[i2];
int id = p.getID();
long type = p.getType();
Object value = p.getValue();
out(" Property ID: " + id + ", type: " + type +
", value: " + value);
}</pre>
<a name="The+Section%27s+Format+ID"></a>
<div class="h2">
<h2>The Section's Format ID</h2>
</div>
<p>The first method called on the <span class="codefrag">Section</span> instance is
<span class="codefrag">getFormatID()</span>. As explained above, the format ID of the
first section in a property set determines the type of the property
set. Its type is <span class="codefrag">ClassID</span> which is essentially a sequence of
16 bytes. A real application using its own type of a custom property set
should have defined a unique format ID and, when reading a property set
stream, should check the format ID is equal to that unique format ID. The
sample program just prints the format ID it finds in a section:</p>
<pre class="code">String s = hex(sec.getFormatID().getBytes());
s = s.substring(0, s.length() - 1);
out(" Format ID: " + s);</pre>
<p>As you can see, the <span class="codefrag">getFormatID()</span> method returns a
<span class="codefrag">ClassID</span> object. An array containing the bytes can be
retrieved with <span class="codefrag">ClassID.getBytes()</span>. In order to get a nicely
formatted printout, the sample program uses the <span class="codefrag">hex()</span> helper
method which in turn uses the POI utility class <span class="codefrag">HexDump</span> in
the <span class="codefrag">org.apache.poi.util</span> package. Another helper method is
<span class="codefrag">out()</span> which just saves typing
<span class="codefrag">System.out.println()</span>.</p>
<a name="The+Properties"></a>
<div class="h2">
<h2>The Properties</h2>
</div>
<p>Before getting the properties, it is possible to find out how many
properties are available in the section via the
<span class="codefrag">Section.getPropertyCount()</span>. The sample application uses this
method to print the number of properties to the standard output:</p>
<pre class="code">int propertyCount = sec.getPropertyCount();
out(" No. of properties: " + propertyCount);</pre>
<p>Now its time to get to the properties themselves. You can retrieve a
section's properties with the method
<span class="codefrag">Section.getProperties()</span>:</p>
<pre class="code">Property[] properties = sec.getProperties();</pre>
<p>As you can see the result is an array of <span class="codefrag">Property</span>
objects. This class has three methods to retrieve a property's ID, its
type, and its value. The following code snippet shows how to call
them:</p>
<pre class="code">for (int i2 = 0; i2 < properties.length; i2++)
{
/* Print a single property: */
Property p = properties[i2];
int id = p.getID();
long type = p.getType();
Object value = p.getValue();
out(" Property ID: " + id + ", type: " + type +
", value: " + value);
}</pre>
<a name="Sample+Output"></a>
<div class="h2">
<h2>Sample Output</h2>
</div>
<p>The output of the sample program might look like the following. It
shows the summary information and the document summary information
property sets of a Microsoft Word document. However, unlike the first and
second section of this HOW-TO the application does not have any code
which is specific to the <span class="codefrag">SummaryInformation</span> and
<span class="codefrag">DocumentSummaryInformation</span> classes.</p>
<pre class="code">Property set stream "/SummaryInformation":
No. of sections: 1
Section 0:
Format ID: 00000000 F2 9F 85 E0 4F F9 10 68 AB 91 08 00 2B 27 B3 D9 ....O..h....+'..
No. of properties: 17
Property ID: 1, type: 2, value: 1252
Property ID: 2, type: 30, value: Titel
Property ID: 3, type: 30, value: Thema
Property ID: 4, type: 30, value: Rainer Klute (Autor)
Property ID: 5, type: 30, value: Test (Stichwörter)
Property ID: 6, type: 30, value: This is a document for testing HPSF
Property ID: 7, type: 30, value: Normal.dot
Property ID: 8, type: 30, value: Unknown User
Property ID: 9, type: 30, value: 3
Property ID: 18, type: 30, value: Microsoft Word 9.0
Property ID: 12, type: 64, value: Mon Jan 01 00:59:25 CET 1601
Property ID: 13, type: 64, value: Thu Jul 18 16:22:00 CEST 2002
Property ID: 14, type: 3, value: 1
Property ID: 15, type: 3, value: 20
Property ID: 16, type: 3, value: 93
Property ID: 19, type: 3, value: 0
Property ID: 17, type: 71, value: [B@13582d
Property set stream "/DocumentSummaryInformation":
No. of sections: 2
Section 0:
Format ID: 00000000 D5 CD D5 02 2E 9C 10 1B 93 97 08 00 2B 2C F9 AE ............+,..
No. of properties: 14
Property ID: 1, type: 2, value: 1252
Property ID: 2, type: 30, value: Test
Property ID: 14, type: 30, value: Rainer Klute (Manager)
Property ID: 15, type: 30, value: Rainer Klute IT-Consulting GmbH
Property ID: 5, type: 3, value: 3
Property ID: 6, type: 3, value: 2
Property ID: 17, type: 3, value: 111
Property ID: 23, type: 3, value: 592636
Property ID: 11, type: 11, value: false
Property ID: 16, type: 11, value: false
Property ID: 19, type: 11, value: false
Property ID: 22, type: 11, value: false
Property ID: 13, type: 4126, value: [B@56a499
Property ID: 12, type: 4108, value: [B@506411
Section 1:
Format ID: 00000000 D5 CD D5 05 2E 9C 10 1B 93 97 08 00 2B 2C F9 AE ............+,..
No. of properties: 7
Property ID: 0, type: 0, value: {6=Test-JaNein, 5=Test-Zahl, 4=Test-Datum, 3=Test-Text, 2=_PID_LINKBASE}
Property ID: 1, type: 2, value: 1252
Property ID: 2, type: 65, value: [B@c9ba38
Property ID: 3, type: 30, value: This is some text.
Property ID: 4, type: 64, value: Wed Jul 17 00:00:00 CEST 2002
Property ID: 5, type: 3, value: 27
Property ID: 6, type: 11, value: true
No property set stream: "/WordDocument"
No property set stream: "/CompObj"
No property set stream: "/1Table"</pre>
<p>There are some interesting items to note:</p>
<ul>
<li>The first property set (summary information) consists of a single
section, the second property set (document summary information) consists
of two sections.</li>
<li>Each section type (identified by its format ID) has its own domain of
property ID. For example, in the second property set the properties with
ID 2 have different meanings in the two section. By the way, the format
IDs of these sections are <strong>not</strong> equal, but you have to
look hard to find the difference.</li>
<li>The properties are not in any particular order in the section,
although they slightly tend to be sorted by their IDs.</li>
</ul>
<a name="Property+IDs"></a>
<div class="h2">
<h2>Property IDs</h2>
</div>
<p>Properties in the same section are distinguished by their IDs. This is
similar to variables in a programming language like Java, which are
distinguished by their names. But unlike variable names, property IDs are
simple integral numbers. There is another similarity, however. Just like
a Java variable has a certain scope (e.g. a member variables in a class),
a property ID also has its scope of validity: the section.</p>
<p>Two property IDs in sections with different section format IDs
don't have the same meaning even though their IDs might be equal. For
example, ID 4 in the first (and only) section of a summary
information property set denotes the document's author, while ID 4 in the
first section of the document summary information property set means the
document's byte count. The sample output above does not show a property
with an ID of 4 in the first section of the document summary information
property set. That means that the document does not have a byte
count. However, there is a property with an ID of 4 in the
<em>second</em> section: This is a user-defined property ID - we'll get
to that topic in a minute.</p>
<p>So, how can you find out what the meaning of a certain property ID in
the summary information and the document summary information property set
is? The standard property sets as such don't have any hints about the
<strong>meanings of their property IDs</strong>. For example, the summary
information property set does not tell you that the property ID 4 stands
for the document's author. This is external knowledge. Microsoft defined
standard meanings for some of the property IDs in the summary information
and the document summary information property sets. As a help to the Java
and POI programmer, the class <span class="codefrag">PropertyIDMap</span> in the
<span class="codefrag">org.apache.poi.hpsf.wellknown</span> package defines constants
for the "well-known" property IDs. For example, there is the
definition</p>
<pre class="code">public final static int PID_AUTHOR = 4;</pre>
<p>These definitions allow you to use symbolic names instead of
numbers.</p>
<p>In order to provide support for the other way, too, - i.e. to map
property IDs to property names - the class <span class="codefrag">PropertyIDMap</span>
defines two static methods:
<span class="codefrag">getSummaryInformationProperties()</span> and
<span class="codefrag">getDocumentSummaryInformationProperties()</span>. Both return
<span class="codefrag">java.util.Map</span> objects which map property IDs to
strings. Such a string gives a hint about the property's meaning. For
example,
<span class="codefrag">PropertyIDMap.getSummaryInformationProperties().get(4)</span>
returns the string "PID_AUTHOR". An application could use this string as
a key to a localized string which is displayed to the user, e.g. "Author"
in English or "Verfasser" in German. HPSF might provide such
language-dependend ("localized") mappings in a later release.</p>
<p>Usually you won't have to deal with those two maps. Instead you should
call the <span class="codefrag">Section.getPIDString(int)</span> method. It returns the
string associated with the specified property ID in the context of the
<span class="codefrag">Section</span> object.</p>
<p>Above you learned that property IDs have a meaning in the scope of a
section only. However, there are two exceptions to the rule: The property
IDs 0 and 1 have a fixed meaning in <strong>all</strong> sections:</p>
<table class="ForrestTable" cellspacing="1" cellpadding="4">
<tr class="b">
<th colspan="1" rowspan="1">Property ID</th>
<th colspan="1" rowspan="1">Meaning</th>
</tr>
<tr class="a">
<td colspan="1" rowspan="1">0</td>
<td colspan="1" rowspan="1">The property's value is a <strong>dictionary</strong>, i.e. a
mapping from property IDs to strings.</td>
</tr>
<tr class="b">
<td colspan="1" rowspan="1">1</td>
<td colspan="1" rowspan="1">The property's value is the number of a <strong>codepage</strong>,
i.e. a mapping from character codes to characters. All strings in the
section containing this property must be interpreted using this
codepage. Typical property values are 1252 (8-bit "western" characters,
ISO-8859-1), 1200 (16-bit Unicode characters, UFT-16), or 65001 (8-bit
Unicode characters, UFT-8).</td>
</tr>
</table>
<a name="Property+types"></a>
<div class="h2">
<h2>Property types</h2>
</div>
<p>A property is nothing without its value. It is stored in a property set
stream as a sequence of bytes. You must know the property's
<strong>type</strong> in order to properly interpret those bytes and
reasonably handle the value. A property's type is one of the so-called
Microsoft-defined <strong>"variant types"</strong>. When you call
<span class="codefrag">Property.getType()</span> you'll get a <span class="codefrag">long</span> value
which denoting the property's variant type. The class
<span class="codefrag">Variant</span> in the <span class="codefrag">org.apache.poi.hpsf</span> package
holds most of those <span class="codefrag">long</span> values as named constants. For
example, the constant <span class="codefrag">VT_I4 = 3</span> means a signed integer value
of four bytes. Examples of other types are <span class="codefrag">VT_LPSTR = 30</span>
meaning a null-terminated string of 8-bit characters, <span class="codefrag">VT_LPWSTR =
31</span> which means a null-terminated Unicode string, or <span class="codefrag">VT_BOOL
= 11</span> denoting a boolean value.</p>
<p>In most cases you won't need a property's type because HPSF does all
the work for you.</p>
<a name="Property+values"></a>
<div class="h2">
<h2>Property values</h2>
</div>
<p>When an application wants to retrieve a property's value and calls
<span class="codefrag">Property.getValue()</span>, HPSF has to interpret the bytes making
out the value according to the property's type. The type determines how
many bytes the value consists of and what
to do with them. For example, if the type is <span class="codefrag">VT_I4</span>, HPSF
knows that the value is four bytes long and that these bytes
comprise a signed integer value in the little-endian format. This is
quite different from e.g. a type of <span class="codefrag">VT_LPWSTR</span>. In this case
HPSF has to scan the value bytes for a Unicode null character and collect
everything from the beginning to that null character as a Unicode
string.</p>
<p>The good new is that HPSF does another job for you, too: It maps the
variant type to an adequate Java type.</p>
<table class="ForrestTable" cellspacing="1" cellpadding="4">
<tr class="b">
<th colspan="1" rowspan="1">Variant type:</th>
<th colspan="1" rowspan="1">Java type:</th>
</tr>
<tr class="a">
<td colspan="1" rowspan="1">VT_I2</td>
<td colspan="1" rowspan="1">java.lang.Integer</td>
</tr>
<tr class="b">
<td colspan="1" rowspan="1">VT_I4</td>
<td colspan="1" rowspan="1">java.lang.Long</td>
</tr>
<tr class="a">
<td colspan="1" rowspan="1">VT_FILETIME</td>
<td colspan="1" rowspan="1">java.util.Date</td>
</tr>
<tr class="b">
<td colspan="1" rowspan="1">VT_LPSTR</td>
<td colspan="1" rowspan="1">java.lang.String</td>
</tr>
<tr class="a">
<td colspan="1" rowspan="1">VT_LPWSTR</td>
<td colspan="1" rowspan="1">java.lang.String</td>
</tr>
<tr class="b">
<td colspan="1" rowspan="1">VT_CF</td>
<td colspan="1" rowspan="1">byte[]</td>
</tr>
<tr class="a">
<td colspan="1" rowspan="1">VT_BOOL</td>
<td colspan="1" rowspan="1">java.lang.Boolean</td>
</tr>
</table>
<p>The bad news is that there are still a couple of variant types HPSF
does not yet support. If it encounters one of these types it
returns the property's value as a byte array and leaves it to be
interpreted by the application.</p>
<p>An application retrieves a property's value by calling the
<span class="codefrag">Property.getValue()</span> method. This method's return type is the
abstract <span class="codefrag">Object</span> class. The <span class="codefrag">getValue()</span> method
looks up the property's variant type, reads the property's value bytes,
creates an instance of an adequate Java type, assigns it the property's
value and returns it. Primitive types like <span class="codefrag">int</span> or
<span class="codefrag">long</span> will be returned as the corresponding class,
e.g. <span class="codefrag">Integer</span> or <span class="codefrag">Long</span>.</p>
<a name="Dictionaries"></a>
<div class="h2">
<h2>Dictionaries</h2>
</div>
<p>The property with ID 0 has a very special meaning: It is a
<strong>dictionary</strong> mapping property IDs to property names. We
have seen already that the meanings of standard properties in the
summary information and the document summary information property sets
have been defined by Microsoft. The advantage is that the labels of
properties like "Author" or "Title" don't have to be stored in the
property set. However, a user can define custom fields in, say, Microsoft
Word. For each field the user has to specify a name, a type, and a
value.</p>
<p>The names of the custom-defined fields (i.e. the property names) are
stored in the document summary information second section's
<strong>dictionary</strong>. The dictionary is a map which associates
property IDs with property names.</p>
<p>The method <span class="codefrag">Section.getPIDString(int)</span> not only returns with
the well-known property names of the summary information and document
summary information property sets, but with self-defined properties,
too. It should also work with self-defined properties in self-defined
sections.</p>
<a name="Codepage+support"></a>
<div class="h2">
<h2>Codepage support</h2>
</div>
<p>The property with ID 1 holds the number of the codepage which was used
to encode the strings in this section. If this property is not available
in a section, the platform's default character encoding will be
used. This works fine as long as the document being read has been written
on a platform with the same default character encoding. However, if you
receive a document from another region of the world and the codepage is
undefined, you are in trouble.</p>
<p>HPSF's codepage support is only as good as the character encoding
support of the Java Virtual Machine (JVM) the application runs on. If
HPSF encounters a codepage number it assumes that the JVM has a character
encoding with a corresponding name. For example, if the codepage is 1252,
HPSF uses the character encoding "cp1252" to read or write strings. If
the JVM does not have that character encoding installed or if the
codepage number is illegal, an UnsupportedEncodingException will be
thrown. This works quite well with Java 2 Standard Edition (J2SE)
versions since 1.4. However, under J2SE 1.3 or lower you are out of
luck. You should install a newer J2SE version to process codepages with
HPSF.</p>
<p>There are some exceptions to the rule saying that a character
encoding's name is derived from the codepage number by prepending the
string "cp" to it. In these cases the codepage number is mapped to a
well-known character encoding name. Here are a few examples:</p>
<dl>
<dt>Codepage 932</dt>
<dd>is mapped to the character encoding "SJIS".</dd>
<dt>Codepage 1200</dt>
<dd>is mapped to the character encoding "UTF-16".</dd>
<dt>Codepage 65001</dt>
<dd>is mapped to the character encoding "UTF-8".</dd>
</dl>
<p>More of these mappings between codepage and character encoding name are
hard-coded in the classes <span class="codefrag">org.apache.poi.hpsf.Constants</span> and
<span class="codefrag">org.apache.poi.hpsf.VariantSupport</span>. Probably there will be a
need to add more mappings. The HPSF author will appreciate any hints.</p>
<a name="sec5"></a>
<a name="Writing+Properties"></a>
<div class="h4">
<h4>Writing Properties</h4>
</div>
<div class="frame note">
<div class="label">Note</div>
<div class="content">This section describes how to write properties.</div>
</div>
<a name="Overview+of+Writing+Properties"></a>
<div class="h2">
<h2>Overview of Writing Properties</h2>
</div>
<p>Writing properties is possible at a high level and at a low level:</p>
<ul>
<li>Most users will want to create or change entries in the summary
information or document summary information streams. </li>
<li>On the low level, there are no convenience classes or methods. You
have to deal with things like property IDs and variant types to write
properties. Therefore you should have read <a href="#sec3">section
3</a> to understand the description of the low-level writing
functions.</li>
</ul>
<p>HPSF's writing capabilities come with the classes
<span class="codefrag">MutablePropertySet</span>, <span class="codefrag">MutableSection</span>,
<span class="codefrag">MutableProperty</span>, and some helper classes. The "mutable"
classes extend their respective superclasses <span class="codefrag">PropertySet</span>,
<span class="codefrag">Section</span>, and <span class="codefrag">Property</span> and provide "set" and
"write" methods, following the <a href="http://en.wikipedia.org/wiki/Decorator_pattern">Decorator
pattern</a>.</p>
<a name="Low-Level+Writing%3A+An+Overview"></a>
<div class="h2">
<h2>Low-Level Writing: An Overview</h2>
</div>
<p>When you are going to write a property set stream your application has
to perform the following steps:</p>
<ol>
<li>Create a <span class="codefrag">MutablePropertySet</span> instance.</li>
<li>Get hold of a <span class="codefrag">MutableSection</span>. You can either retrieve
the one that is always present in a new <span class="codefrag">MutablePropertySet</span>,
or you have to create a new <span class="codefrag">MutableSection</span> and add it to
the <span class="codefrag">MutablePropertySet</span>.
</li>
<li>Set any <span class="codefrag">Section</span> fields as you like.</li>
<li>Create as many <span class="codefrag">MutableProperty</span> objects as you need. Set
each property's ID, type, and value. Add the
<span class="codefrag">MutableProperty</span> objects to the
<span class="codefrag">MutableSection</span>.
</li>
<li>Create further <span class="codefrag">MutableSection</span>s if you need them.</li>
<li>Eventually retrieve the property set as a byte stream using
<span class="codefrag">MutablePropertySet.toInputStream()</span> and write it to a POIFS
document.</li>
</ol>
<a name="Low-level+Writing+Functions+In+Details"></a>
<div class="h2">
<h2>Low-level Writing Functions In Details</h2>
</div>
<p>Writing properties is introduced by an artificial but simple example: a
program creating a new document (aka POI file system) which contains only
a single document: a summary information property set stream. The latter
will hold the document's title only. This is artificial in that it does
not contain any Word, Excel or other kind of useful application document
data. A document containing just a property set is without any practical
use. However, it is perfectly fine for an example because it make it very
simple and easy to understand, and you will get used to writing
properties in real applications quickly.</p>
<p>The application expects the name of the POI file system to be written
on the command line. The title property it writes is "Sample title".</p>
<p>Here's the application's source code. You can also find it in the
"examples" section of the POI source code distribution. Explanations are
following below.</p>
<pre class="code">package org.apache.poi.hpsf.examples;
import java.io.FileOutputStream;
import java.io.IOException;
import java.io.InputStream;
import org.apache.poi.hpsf.MutableProperty;
import org.apache.poi.hpsf.MutablePropertySet;
import org.apache.poi.hpsf.MutableSection;
import org.apache.poi.hpsf.SummaryInformation;
import org.apache.poi.hpsf.Variant;
import org.apache.poi.hpsf.WritingNotSupportedException;
import org.apache.poi.hpsf.wellknown.PropertyIDMap;
import org.apache.poi.hpsf.wellknown.SectionIDMap;
import org.apache.poi.poifs.filesystem.POIFSFileSystem;
/**
* <p>This class is a simple sample application showing how to create a property
* set and write it to disk.</p>
*
* @author Rainer Klute
* @since 2003-09-12
*/
public class WriteTitle
{
/**
* <p>Runs the example program.</p>
*
* @param args Command-line arguments. The first and only command-line
* argument is the name of the POI file system to create.
* @throws IOException if any I/O exception occurs.
* @throws WritingNotSupportedException if HPSF does not (yet) support
* writing a certain property type.
*/
public static void main(final String[] args)
throws WritingNotSupportedException, IOException
{
/* Check whether we have exactly one command-line argument. */
if (args.length != 1)
{
System.err.println("Usage: " + WriteTitle.class.getName() +
"destinationPOIFS");
System.exit(1);
}
final String fileName = args[0];
/* Create a mutable property set. Initially it contains a single section
* with no properties. */
final MutablePropertySet mps = new MutablePropertySet();
/* Retrieve the section the property set already contains. */
final MutableSection ms = (MutableSection) mps.getSections().get(0);
/* Turn the property set into a summary information property. This is
* done by setting the format ID of its first section to
* SectionIDMap.SUMMARY_INFORMATION_ID. */
ms.setFormatID(SectionIDMap.SUMMARY_INFORMATION_ID);
/* Create an empty property. */
final MutableProperty p = new MutableProperty();
/* Fill the property with appropriate settings so that it specifies the
* document's title. */
p.setID(PropertyIDMap.PID_TITLE);
p.setType(Variant.VT_LPWSTR);
p.setValue("Sample title");
/* Place the property into the section. */
ms.setProperty(p);
/* Create the POI file system the property set is to be written to. */
final POIFSFileSystem poiFs = new POIFSFileSystem();
/* For writing the property set into a POI file system it has to be
* handed over to the POIFS.createDocument() method as an input stream
* which produces the bytes making out the property set stream. */
final InputStream is = mps.toInputStream();
/* Create the summary information property set in the POI file
* system. It is given the default name most (if not all) summary
* information property sets have. */
poiFs.createDocument(is, SummaryInformation.DEFAULT_STREAM_NAME);
/* Write the whole POI file system to a disk file. */
poiFs.writeFilesystem(new FileOutputStream(fileName));
}
}</pre>
<p>The application first checks that there is exactly one single argument
on the command line: the name of the file to write. If this single
argument is present, the application stores it in the
<span class="codefrag">fileName</span> variable. It will be used in the end when the POI
file system is written to a disk file.</p>
<pre class="code">if (args.length != 1)
{
System.err.println("Usage: " + WriteTitle.class.getName() +
"destinationPOIFS");
System.exit(1);
}
final String fileName = args[0];</pre>
<p>Let's create a property set now. We cannot use the
<span class="codefrag">PropertySet</span> class, because it is read-only. It does not have
a constructor creating an empty property set, and it does not have any
methods to modify its contents, i.e. to write sections containing
properties into it.</p>
<p>The class to use is <span class="codefrag">MutablePropertySet</span>. It is a subclass
of <span class="codefrag">PropertySet</span>. The sample application calls its no-args
constructor in order to establish an empty property set:</p>
<pre class="code">final MutablePropertySet mps = new MutablePropertySet();</pre>
<p>As said, we have an empty property set now. Later we will put some
contents into it.</p>
<p>By the way, the <span class="codefrag">MutablePropertySet</span> class has another
constructor taking a <span class="codefrag">PropertySet</span> as parameter. It creates a
mutable deep copy of the property set given to it.</p>
<p>The <span class="codefrag">MutablePropertySet</span> created by the no-args constructor
is not really empty: It contains a single section without properties. We
can either retrieve that section and fill it with properties or we can
replace it by another section. We can also add further sections to the
property set. The sample application decides to retrieve the section
being already there:</p>
<pre class="code">final MutableSection ms = (MutableSection) mps.getSections().get(0);</pre>
<p>The <span class="codefrag">getSections()</span> method returns the property set's
sections as a list, i.e. an instance of
<span class="codefrag">java.util.List</span>. Calling <span class="codefrag">get(0)</span> returns the
list's first (or zeroth, if you prefer) element. The <span class="codefrag">Section</span>
returned is a <span class="codefrag">MutableSection</span>: a subclass of
<span class="codefrag">Section</span> you can modify.</p>
<p>The alternative to retrieving the <span class="codefrag">MutableSection</span> being
already there would have been to create an new
<span class="codefrag">MutableSection</span> like this:</p>
<pre class="code">MutableSection s = new MutableSection();</pre>
<p>There is also a constructor which takes a <span class="codefrag">Section</span> as
parameter and creates a mutable deep copy of it.</p>
<p>The <span class="codefrag">MutableSection</span> the sample application retrieved from
the <span class="codefrag">MutablePropertySet</span> is still empty. It contains no
properties and does not have a format ID. As you have read <a href="#sec3">above</a> the format ID of the first section in a
property set determines the property set's type. Since our property set
should become a SummaryInformation property set we have to set the format
ID of its first (and only) section to
<span class="codefrag">F29F85E0-4FF9-1068-AB-91-08-00-2B-27-B3-D9</span>. However, you
won't have to remember that ID: HPSF has it defined as the well-known
constant <span class="codefrag">SectionIDMap.SUMMARY_INFORMATION_ID</span>. The sample
application writes it to the section using the
<span class="codefrag">setFormatID(byte[])</span> method:</p>
<pre class="code">ms.setFormatID(SectionIDMap.SUMMARY_INFORMATION_ID);</pre>
<p>Now it is time to create a property. As you might expect there is a
subclass of <span class="codefrag">Property</span> called
<span class="codefrag">MutableProperty</span> with a no-args constructor:</p>
<pre class="code">final MutableProperty p = new MutableProperty();</pre>
<p>A <span class="codefrag">MutableProperty</span> object must have an ID, a type, and a
value (see <a href="#sec3">above</a> for details). The class
provides methods to set these attributes:</p>
<pre class="code">p.setID(PropertyIDMap.PID_TITLE);
p.setType(Variant.VT_LPWSTR);
p.setValue("Sample title");</pre>
<p>The <span class="codefrag">MutableProperty</span> class has a constructor which you can
use to pass in all three attributes in a single call. See the Javadoc API
documentation for details!</p>
<p>The sample property set is complete now. We have a
<span class="codefrag">MutablePropertySet</span> containing a <span class="codefrag">MutableSection</span>
containing a <span class="codefrag">MutableProperty</span>. Of course we could have added
more sections to the property set and more properties to the sections but
we wanted to keep things simple.</p>
<p>The property set has to be written to a POI file system. The following
statement creates it.</p>
<pre class="code">final POIFSFileSystem poiFs = new POIFSFileSystem();</pre>
<p>Writing the property set includes the step of converting it into a
sequence of bytes. The <span class="codefrag">MutablePropertySet</span> class has the
method <span class="codefrag">toInputStream()</span> for this purpose. It returns the
bytes making out the property set stream as an
<span class="codefrag">InputStream</span>:</p>
<pre class="code">final InputStream is = mps.toInputStream();</pre>
<p>If you'd read from this input stream you'd receive all the property
set's bytes. However, it is very likely that you'll never do
that. Instead you'll pass the input stream to the
<span class="codefrag">POIFSFileSystem.createDocument()</span> method, like this:</p>
<pre class="code">poiFs.createDocument(is, SummaryInformation.DEFAULT_STREAM_NAME);</pre>
<p>Besides the <span class="codefrag">InputStream</span> <span class="codefrag">createDocument()</span>
takes a second parameter: the name of the document to be created. For a
SummaryInformation property set stream the default name is available as
the constant <span class="codefrag">SummaryInformation.DEFAULT_STREAM_NAME</span>.</p>
<p>The last step is to write the POI file system to a disk file:</p>
<pre class="code">poiFs.writeFilesystem(new FileOutputStream(fileName));</pre>
<a name="Further+Reading"></a>
<div class="h4">
<h4>Further Reading</h4>
</div>
<p>There are still some aspects of HSPF left which are not covered by this
HOW-TO. You should dig into the Javadoc API documentation to learn
further details. Since you've struggled through this document up to this
point, you are well prepared.</p>
<div id="authors" align="right">by Rainer Klute</div>
</div>
</div>
</div>
</td>
<!--================= end Content ==================-->
</tr>
</tbody>
</table>
<!--================= end Main ==================-->
<!--================= start Footer ==================-->
<div id="footer">
<table summary="footer" cellspacing="0" cellpadding="4" width="100%" border="0">
<tbody>
<tr>
<!--================= start Copyright ==================-->
<td colspan="2">
<div align="center">
<div class="copyright">
Copyright © 2002-2012 The Apache Software Foundation. All rights reserved.<br>
Apache POI, POI, Apache, the Apache feather logo, and the Apache
POI project logo are trademarks of The Apache Software Foundation.
</div>
</div>
</td>
<!--================= end Copyright ==================-->
</tr>
<tr>
<td align="left">
<!--================= start Host ==================-->
<!--================= end Host ==================--></td><td align="right">
<!--================= start Credits ==================-->
<div align="right">
<div class="credit"></div>
</div>
<!--================= end Credits ==================-->
</td>
</tr>
</tbody>
</table>
</div>
<!--================= end Footer ==================-->
</body>
</html>
