May 12 2014

SQL Server Reporting Services (SSRS) emits dodgy Word docx documents

By now we’re used to products which emit docx files which are umm, not .. quite .. right.

But its more noteworthy when the product in question is from Microsoft.  After all, its their file format (ECMA etc standardisation notwithstanding).

The product in question here is SQL Server Reporting Services 2012 and its Word export.

It seems they didn’t bother to validate their documents (eg using Open XML SDK 2.0 Productivity Tool):

Apparently there’s a reason for this:

“Word and SSRS treat page headers and footers differently. Word actually positions them inside the page margins, whereas SSRS positions them inside the area that the margins surround. As a result, in Word, the page margins do not control the distance between the top edge of the page and that of the page header (or similarly for the page footer). Instead, Word has separate “Header from Top” and “Footer from Bottom” properties to control those distances. Since RDL does not have equivalent properties, the Word renderer sets these properties to zero.”
But the problem is that it is actually setting them to blank (as opposed to zero), which is not valid.

Another problem:

JAXB doesn’t like invalid documents, so docx4j has to fix these sorts of things before it can construct a content model.  (Maybe that’s why SSRS calls it Word export, not docx export:- they just check Word can open the document, then call it job done)

There are other problems with SSRS docx which the Productivity Tool doesn’t report.

Take a look at the styles part:

Notice anything wrong?  It’d be better if the EmptyCellLayoutStyle had @w:styleId and @w:type, like so:

It’d also be nice if it defined the “Normal” style it is basedOn!

docx4j and other consumers could/should detect such problems and degrade gracefully in the face of them, but Microsoft (of all companies!) should exercise better quality control.

Mar 16 2014

docx4j and Google Drive

Given the news this week about Google lowering prices per GB on Google Drive, I thought it would be timely to explore interop with docx4j. is a small project which demonstrates:

Clone the project, and set it up using Maven in your IDE.  I’m not going to tell you how to do that.
Enabling the Drive API
From there, it is fairly straightforward  (assuming you have a Google account); you just need to enable the Drive API: set up a project and application in the Developers Console:
  • press the red “CREATE NEW CLIENT ID” button, then choose application type “Installed Application”; I then chose subtype “Other”
  • hit the “Download JSON” button; save it as client_secret.json in your project dir

Run our code

OK, now try running Docx4jUploadToGoogleDrive

It ought to say something like:

Please open the following URL in your browser then type the authorization code:…

Paste the auth code into your IDE’s console (, probably the same place which displayed the above message) then press enter.  If you aren’t logged into your Google account in your browser, its at this point that you’ll be asked to log in.

The code will create a new docx file, and after uploading it, if successful, report the File ID allocated by Google Drive:

File ID: 0CyHdofN18p16OF9YWWNFUFdmTjg

The other 2 samples require you to provide an auth code the same way (each time you run them).  Obviously, you’d be more sophisticated than this in a production application.  See further

Nov 28 2013

docx4j 3.0 and Maven

blog/2011/10/hello-maven-central/ walks you through the basics of using docx4j in an Eclipse project with the help of m2eclipse.

This post is about the different ways you can set up docx4j 3.0 with the help of Maven.

We’ll be using the following skeleton pom.xml:

<?xml version="1.0" encoding="UTF-8"?>
<project xmlns="" xmlns:xsi="" xsi:schemaLocation="">

		some description




		<!-- dependencies go here -->



Adding the core dependency

To use docx4j, including its LGPL XHTML import capability, just include the following dependency in your pom.xml:

That’ll drag in docx4j, and all the other dependencies (you should be able to see then in Eclipse under Maven Dependencies, or by running mvn dependency:tree at a command prompt).
If you don’t want the XHTML import stuff, just use:

(You should consider adding a to your classpath)
Both of the above default to using log4j.  If you are happy with log4j, you’ll want a log4j.xml file unless you already have it on your classpath.  If you don’t, you can configure to suit.
If you want to use something other than log4j for logging, well you can, since docx4j uses slf4j.
First you need to exclude the log4j stuff.


Then you add in the dependencies for your other logging frameworks.   See further and slf4j in
docx4j relies very heavily on JAXB.  With Java 6 or 7, usually it’ll use the JAXB included in that (though things can be different with application servers – see the deployment forums for details).
The point here is that there is an alternative JAXB implementation, called EclipseLink MOXy (see, which is very well supported by its developers.  You can try it with docx4j.  To do so, just include the following additional dependencies:




Since using MOXy with docx4j is all quite new, you may run into some minor issues.  If you do, please let us know in the docx4j forums (with sufficient info for us to reproduce what you are seeing!).  Thanks.

Nov 26 2013

docx4j 3.0 released

On behalf of everyone who has contributed to docx4j, Plutext is pleased to announce that version 3 was released today.

You can get it from Maven Central, or from (the jar, the dependencies, or everything including documentation zipped up)

Source code is available at GitHub or from the Maven Central link above.  Javadoc is at Maven Central.

For what you need to know about docx4j 3.0, please see this post.

The XHTML Import stuff is now a separate project (since it and its dependencies are LGPL, not ASLv2 like docx4j).

  • the three jars you need (docx4j-ImportXHTML, xhtmlrenderer, and iText) are included for convenience in the zip file above.  You can delete them if you don’t need or want XHTML import.
  • or you can get it from Maven Central

docx4j 3.0 uses slf4j for logging.  For convenience, log4j is the default implementation.  A follow-up post will explain more about logging config.

Thanks to everyone who has helped to make this release our best yet!

If you have questions pertaining to the use of docx4j, please post them in our forum, or on StackOverflow (rather than in comments to this post).

Nov 07 2013

docx4j 3.0 beta

A beta of docx4j 3.0 is now available, at: [link updated 15 Nov]

That zip file contains docx4j, and all its dependencies.  To use it, add all the jars to your classpath.

Alternatively, Maven users can get the beta from our staging repo on GitHub.


docx4j 3.0 beta is:


Our last blog post outlines the major things to be aware of in v3.

Additional notes:

  • For convenience, the zip file also contains docx4j-ImportXHTML, and its dependencies, which are LGPL.  You can delete these if you wish.  They aren’t in the mvn staging repo.
  • To see any logging, you’ll need to add an slf4j implementation.
  • You might want to add a file

You can find updated Getting Started guide in docx|pdf formats at

Feedback welcome.  You can reply here, or to the post in the docx4j forums.

All going smoothly, we’ll progress to final release over the next couple of weeks, so the sooner your feedback, the better!

Oct 18 2013

docx4j 3.0 – what you need to know

docx4j 3.0 (beta for which will be available shortly) contains a lot of changes, some big, some small.

Here are the most visible (see our changelog for the rest):


docx4j 3.0 uses slf4j, instead of log4j.

As the slf4j website puts it:

The Simple Logging Facade for Java (SLF4J) serves as a simple facade or abstraction for various logging frameworks (e.g. java.util.logging, logback, log4j) allowing the end user to plug in the desired logging framework at deployment time.

So you need the slf4j api jar on your classpath:


If you want to use log4j, then include it, and:


XHTML Import

The XHTML Import functionality is now a separate project on GitHub.

The reason being that its main dependency – Flying Saucer - is licensed under LGPL v2.1 (as opposed to ASL v2, which docx4j’s other dependencies use).

If you want this functionality, you have to add these jars to your classpath.  We’ll update this post with their coordinates once they are in Maven Central.

Docx4j facade

3.0 contains a facade providing clean access to some typical uses of docx4j:
  • Loading a document
  • Saving a document
  • Binding xml to content controls in a document
  • Exporting the document (to HTML, or PDF and other formats supported by the FO renderer)

You don’t have to use this – in that existing code should continue to work – but the facade is the right way to do things.  Behind the facade is a major rethink/cleanup to the export architecture/implementation, contributed by Alberto.


The key technology underlying docx4j – and a major differentiator from Apache POI – is JAXB.

There is a JAXB reference implementation; the JAXB baked into Java 6 and 7 is based on it.

Prior to v3, you had to use the reference implementation, or the implementation included in the JDK.

With v3, you can choose to use EclipseLink MOXy instead.  To do so, simply include docx4j-MOXy-JAXBContext-3.0.0.jar and the MOXy jars on your classpath.

Sample code

The docx4j samples have relocated to src/samples

May 15 2013

docx4j in a single page

Here’s a single A4 page reference/overview of docx4j aka a cheat sheet, in PDF or PNG format.

This one is focused on docx files (WordprocessingML).

I’ll create something similar for pptx and xlsx over coming days.

May 15 2013

docx4j/pptx/xlsx online code generation

Just launched is

You should be able to see it in the menu at the top right of this website (if not, reload the web page…).

There are three things you can do with it right now:

• Explore your docx/pptx/xlsx and its representation in docx4j

• Convert  docx to PDF or XSL FO

• Merge docx files (eg cover letter plus contract) into a single docx, using Plutext’s MergeDocx. Or the same thing for pptx files, using MergePptx.

Here I want to focus on the first of these.

After you’ve uploaded your docx/pptx/xlsx, the first thing you see is like docx4j’s PartsList sample:

Here, I’ll click in the left hand column to look at the main document part, document.xml

When I do that, I see the XML:

No surprises there.

But notice the hyperlinks.  Here I’ll just click on the first w:p.

What you get back, is Java source code to create that complete structure:-

As you can see from the image above, both styles of code (as described in docx4j’s Getting Started document) are produced for you.  With a bit of luck, you can cut/paste either into your IDE (Eclipse or whatever), and just run with it!

To actually see the created object in an Office document, you’ll still need to add the created object to a part.  See Getting Started, or the cheat sheet for how to do that.

I hope this helps you to create/modify your Office documents more efficiently,with docx4j!

Do let us know what you think in the comments, or in docx4j’s forums.

May 24 2012

docx4j 2.8.0 released

I’m pleased to say that docx4j 2.8.0 is now released.

What is docx4j?

docx4j is an open source (Apache v2) library for working with docx, pptx, and xslx files, based around  JAXB.

What’s new?

The headline feature is XHTML import.  docx4j can convert XHTML to Word document content, formatting it based on the CSS.  Images and tables are supported. See the ConvertInXHTMLDocument and ConvertInXHTMLFragment samples.

Where do you get it?

See our downloads page or:

Binaries: You can download a jar alone or a zip with all deps or pick and choose.  If you’re upgrading from 2.7.1, you need the  docx4j jar and:

Source: the source code is on GitHub at; here’s how to setup docx4j source code

Maven: docx4j 2.8.0 is in Maven Central.  Here is a guide to getting started (where it says 2.7.1, just use 2.8.0).

Getting Started

See the “Getting Started” guide, in html docx or pdf flavours.

There is lots of sample code here (freshly reviewed for 2.8.0).


If you are looking for help (and have read the Getting Started Guide :-) ), you can post in our forums, or on Stack Overflow (where there is a docx4j tag).

Thanks to our contributors

A number of contributions have made this release what it is; thanks very much to those who contributed.

Contributors to this release and a more complete list of changes may be found in README.txt

Thanks also to those who have +1′d pages on this website, or tweeted or blogged about docx4j, which is critical to expanding the docx4j community!

May 18 2012

docx4j from GitHub in Eclipse

docx4j is now on GitHub!

This should make it easier for users to maintain their own branches (public or private), and contribute improvements back.

As of now, GitHub is the project’s authoritative version control.  We’re no longer updating the existing svn repository.

Its pretty easy to work with docx4j sources in Eclipse. This post shows you how.

First, make sure you have eGit installed in Eclipse.  Install it from here.  On Windows, it is also useful to have msysgit.  Refer elsewhere for how to set these up. Update: there is a GitHub Windows client now (I haven’t tried it) which apparently includes msysgit.

You also need m2eclipse

Assuming you’ve done all that, setting up the docx4j source code is just a few steps.

But first, be aware there is a difference between cloning and forking.  Cloning gives you a copy of the source code you can work on, but without more, no easy way to contribute changes back.  Forking sets you up with the source code, and makes it easy to contribute changes back.

If you think you might be making changes to the docx4j source code, you’re probably best to create a fork on GitHub right from the start.

Step 1 (optional, but recommended): To create a fork, log in to GitHub, visit then press the “Fork” button.

Step 2: Create your local repository (git clone)

This can be done from within Eclipse, or using Git Gui (easiest), or Git Bash Shell.

To do it from within Eclipse, File > Import .. > Repositories from GitHub:

If you forked docx4j, find your fork (it might not appear immediately, which is why Git Gui or Git Bash Shell are better for this step), select it, and click next.

If you didn’t fork docx4j, type ‘docx4j’ then press ‘search’, the plutext/docxj repository should come up:

Select plutext/docx4j, then click next.

This creates a local git repository on your computer.

Step 3: Now you need to import that repository into Eclipse as a project:

File > Import .. > Projects from Git

Eclipse should find the existing project settings:

(If it didn’t and you had to use the new projects wizard; be sure to set the file location to wherever your git repository is, rather than letting Eclipse create a new empty project in the workspace)

Now you should have a docx4j project in Eclipse, and it should be properly configured (since the project settings come with the project).

You should be done. But if something isn’t right, you can configure it manually (see further below).

Next steps?  Improve the docx4j source code in Eclipse :-) , then Team > Commit, to commit those changes to your local repository.

Made a change which would be useful to others?  If you forked docx4j as per step 1 above, you can push your changes to your repository on GitHub, then send a pull request.

If you didn’t fork docx4j, do that now on GitHub, then configure things locally to push your changes to your repository on GitHub, then you’ll be right to push your changes to your repository on GitHub, then send a pull request.  Other docx4j users will thank you for this :-)

Manual configuration:

Configure > Convert to Maven Project

Properties > Java Compiler > Compiler compliance level: change to 1.6

Java Build Path > Libraries: remove 1.5 system library; Add Library … JRE System Library .. 1.6

Java Build Path > Source: check none of the entries say “Excluded: **” (remove the exclusion)