Introduction
I have been using iTextSharp for sometime now but had not been able to create a dynamic PDF that can display both English and Chinese characters successfully until today. There are not many writings out there that would give me a complete coverage on how this is done, so I figured it will be beneficial to folks who are still struggling like I was that I summarize all the important steps in one post for easier reference on this tricky issue. Even though my case only deals with Simplified Chinese characters, the same technique is applicable to any double-byte Unicode character set, I believe.
There is plenty of documentation about how to use the iTextSharp API to create dynamic PDF, so here I will only focus on what is generally missing in those web posts – getting Chinese or Asian characters in general to display correctly alongside with English ones.
Getting the API
Get the latest version of iTextSharp.dll from here. Once downloaded and unzipped, there is only one file itextsharp.dll; I added a reference to it in my C# project and that was it. I also downloaded iTextAsian.dll and iTextAsianCmaps.dll (from the Extras link on the same download page for iTextSharp.dll) and added references to them, but it turned out I did not need them at all for what I wanted to accomplish, so later I removed them from the project.
Getting the Font File
This was a turning point for me and everything turned out just fine after I downloaded the correct TTF file for the Simplified Chinese characters I wanted to display. There are many font files for the Simplified Chinese language (used mostly in mainland China); my eyes almost dried up after exhaustive Google searches, and I used a trick in Microsoft Word to help me sort this out – I opened Word, switched the input language to Simplified Chinese, and started typing in a few Simplified Chinese letters into the document. When I looked at the Font faces dropdown box on the top-left corner of the menu bar, it automatically changed from the default “Calibri” to “SimSun”, so that settled it – “SimSun.ttf” was what I wanted to get. I then simply typed “simsun.ttf” on the Google search box and picked the top one from the results: http://www.gamefront.com/files/16488629/simsun.ttf/. There is no need to install the font file. I just placed it in a “Fonts” folder (you can name it whatever) inside my ASP.NET project that is accessible by the runtime ASP.NET code.
Generating a Mixed-Language PDF
Now, it’s show time. In this theme, I have an ASP.NET grid view that contains both English and Chinese texts that are pulled from a SQL Server 2008 database. This is a calendar displaying school events and schedules; so it is natural that a “Save as PDF” link is displayed here to allow the user to grab a copy of the school calendar to go. It was quite straightforward to accomplish this, if there were no Chinese characters involved - I wrapped the grid view data into a DataTable
object and passed it to a pre-written function called ExportPdfTable
that in turn made the appropriate iTextSharp API calls to carry out the dirty work. Below is the code snippet that traverses the DataTable
object and spits out a PDF document on the fly:
…
using iTextSharp.text
using iTextSharp.text.pdf;
..
public static bool ExportPdfTable(string PdfFileName, DataTable dt,
string DocTitle,string PageHeader,string HeaderImgPath)
{
Document doc = new Document();
try
{
string physicalFile = PdfFileName;
PdfWriter pw=PdfWriter.GetInstance(doc,
new FileStream(physicalFile, FileMode.Create));
doc.Open();
if (PageHeader.Length > 0)
doc.Add(new Paragraph(new Phrase("")));
if (DocTitle.Length > 0)
doc.Add(new Phrase(DocTitle));
if (dt.Columns.Count == 0) return false;
int cols = dt.Columns.Count;
int rows = dt.Rows.Count;
PdfPTable t = new PdfPTable(cols);
t.WidthPercentage = 100;
PdfPCell c;
string fontpath = System.Web.HttpContext.Current.Request.PhysicalApplicationPath +
"\\includes\\fonts\\simsun.ttf";
BaseFont bf = BaseFont.CreateFont(fontpath,BaseFont.IDENTITY_H,
BaseFont.EMBEDDED);
Font fontContent = new Font(bf, 11);
Font fontHeader = new Font(bf, 12);
for (int j = 0; j < cols; j++)
{
Phrase pr=new Phrase((dt.Columns[j].Caption != null &&
dt.Columns[j].Caption.Length > 0) ? dt.Columns[j].Caption :
dt.Columns[j].ColumnName,fontHeader);
c = new PdfPCell(pr);
c.PaddingBottom = 5f;
c.PaddingTop = 5f;
c.PaddingLeft = 8f;
c.PaddingRight = 8f;
t.AddCell(c);
}
for (int i = 0; i < rows; i++)
{
for (int j = 0; j < cols; j++)
{
c = new PdfPCell(new Phrase(dt.Rows[i][j].ToString(),fontContent));
c.PaddingBottom = 5f;
c.PaddingTop = 5f;
c.PaddingLeft = 8f;
c.PaddingRight = 8f;
t.AddCell(c);
}
}
return doc.Add(t);
}
catch (Exception ex)
{
Console.WriteLine(ex.Message);
return false;
}
finally
{
doc.Close();
}
}
The four lines that are critical for allowing Chinese characters to show up in the dynamically generated PDF are shighlighted here:
string fontpath =
System.Web.HttpContext.Current.Request.PhysicalApplicationPath +
"\\includes\\fonts\\simsun.ttf";
BaseFont bf = BaseFont.CreateFont(fontpath,
BaseFont.IDENTITY_H, BaseFont.EMBEDDED);
Font fontContent = new Font(bf, 11);
Font fontHeader = new Font(bf, 12);
If these four lines are taken out, the function would still output the same good-looking PDF file except that all Chinese characters would disappear completely. The key in creating a correct base font in this case is loading the correct TTF file and setting the encoding to BaseFont.IDENTITY_H
. I tried “UTF-8” and some other values, but none of them seemed to work. I got this tip from a very brief but helpful post at http://stackoverflow.com/questions/1727765/itextsharp-international-text and I am grateful to its author StewBob and to stackoverflow.com that hosted the post.
Deployment
Everything ran perfectly on my local development machine... the PDF file got created and the Chinese characters showed up...but a "That assembly does not allow partially trusted callers." error was thrown after the project was deployed to an external hosting site. What was going on? Well, this was actually not a new problem and had nothing to do with the Chinese font. As it turned out, I was not supposed to just simply upload itextsharp.dll to the hosting server. I needed to download the iTextSharp source files and compile the DLL with some additional security attribute added. So these extra steps were taken to fix this DLL error:
- Downloaded the iTextSharp source files (version 5.xx) from http://sourceforge.net/projects/itextsharp/files/itextsharp/.
- Launched the iTextSharp solution and opened AssemblyInfo.cs and added the following lines to the end of the file:
[assembly: AssemblyDelaySign(false)]
[assembly: AssemblyKeyFile("")]
[assembly: AssemblyKeyName("")]
[assembly: AllowPartiallyTrustedCallers()]
- Compiled the DLL and copied it to the hosting site's bin folder, and done.
Summary
Wrapping it up, the four key steps are:
- Download the correct font file and place it in a folder inside the web project.
- Create a
BaseFont
with the font file loaded and set the encoding to BaseFont.IDENTITY_H
; also set embedding the font in the PDF document to true (BaseFont.EMBEDDED
).
- Create new fonts based on the
BaseFont
and use them in Paragraph
or Phrase
constructors.
- Make sure when iTextSharp.dll is deployed to the remote hosting server, the DLL must be re-compiled with the "
AllowPartiallyTrustedCallers
" attribute set.
History
- First submitted on 5/12/2011.