Inserting PDF Attachments into Salesforce.com using Talend and iText

I recently presented at the local Chicago Force.com Developer Group on using Talend to move data into and out of Salesforce.com. One question I fielded was how about moving documents into Salesforce.com. Yes you can use Talend to do this! For this demonstration I am going to use a few components and show you how to get the class path to execute properly for the ETL job.

Use Case:

For this demonstration, we are going to dynamically generate PDF content inside our ETL job. We are going to extract all our Account records from SFDC and dynamically generate PDF content from the fields on the Account records. We will then inser that data into an Attachment record for each of the Accounts as a PDF document created using iText.

Step 1 - Create Your Salesforce.com Metadata Connection

The first step is to create your connection for SFDC metadata. I have a video on YouTube which shows you how can do this. For this example you will want to pull in the Accounts and Attachements objects from SFDC. After following the instructions in the video for creating your SFDC metadata connections you should have them in your repository like figure below.

Step 2 - Download iText PDF Library

You will need the iText PDF library for the portions of the ETL job which will generate dynamic PDF content. You will also need to add the iText jar file to you java build path. You can do this by using the User Libraries feature as shown below:

Talend -> Preferences -> Java -> Build Path -> User Libraries

Step 3 - Create a new Job Design and Load Java Libraries

Talend allows you to load external Java libraries into your job which can have code executed inside your job. The component to load external Java libraries is called tLibraryLoad. You should use this component as one of the first steps in your job to load the any dependent jars you need. In this case, we are going to load the iText PDF library, as well as Apache Commons Codec so that we can Base64 encode our Attachment file content (more on that later).

Using tLibraryLoad component to load Java jar files.

Step 4 - Query SFDC Account Records

This step we simply use the tSalesforceInput component to read Accounts. For details you can see the video above on how to the Account records.

Step 4 - Create the Account Record PDF using tJavaRow

The tJavaRow component is a very powerful component. It allows you to code functionality into your ETL job using Java. In this example we are using the iText PDF Library to generate a simple PDF. The thing to remember about tJavaRow is that the code will execute for each record in the input step.

When you add and connect your tJavaRow component to the tSalesforceInput component, the first thing you need to do is click on the Advanced Settings tab and add any import for libraries that you need. These classes should reside in the jars you loaded in the previous step inside tLibraryLoad component.

Import Libraries for iText and Apache Commons Codec

Once you have the imports setup like above, you can then go to the Basic Settings tab. This will present you with the tJavaRow code editor where you can add your code. The first thing I do is click "Sync Columns" and "Generate Code" buttons. This will automatically create the code to simply move all the data from the input into your output row for this component.

Auto Code Generation and Column Sync

After I have let Talend do the heavy lifting of generating my getter and setter code, I then add two new fields. One called 'FileName' and one called "Content_Body". These fields will hold the filename in the format "Account Name.pdf" and the actual file content as a Base64 encoded string.

Add Fields to the Mapping for the Content and File Name

 The Salesforce API's use Base64 Strings to encode and pass file content. That is the reason we need to use Apache Commons in our job, to convert the PDF bytes into a format that SFDC and ingest. Using the output and input variable in the code, we can generate a PDF using the simple iText PDF objects. Here is the complete code. All this code goes inside the tJavaRow component.

//Code generated according to input schema and output schema

output_row.Id = input_row.Id;

output_row.IsDeleted = input_row.IsDeleted;

output_row.MasterRecordId = input_row.MasterRecordId;

output_row.Name = input_row.Name;

output_row.Type = input_row.Type;

output_row.ParentId = input_row.ParentId;

output_row.BillingStreet = input_row.BillingStreet;

output_row.BillingCity = input_row.BillingCity;

output_row.BillingState = input_row.BillingState;

output_row.BillingPostalCode = input_row.BillingPostalCode;

output_row.BillingCountry = input_row.BillingCountry;

output_row.ShippingStreet = input_row.ShippingStreet;

output_row.ShippingCity = input_row.ShippingCity;

output_row.ShippingState = input_row.ShippingState;

output_row.ShippingPostalCode = input_row.ShippingPostalCode;

output_row.ShippingCountry = input_row.ShippingCountry;

output_row.Phone = input_row.Phone;

output_row.Fax = input_row.Fax;

output_row.AccountNumber = input_row.AccountNumber;

output_row.Website = input_row.Website;

output_row.Sic = input_row.Sic;

output_row.Industry = input_row.Industry;

output_row.AnnualRevenue = input_row.AnnualRevenue;

output_row.NumberOfEmployees = input_row.NumberOfEmployees;

output_row.Ownership = input_row.Ownership;

output_row.TickerSymbol = input_row.TickerSymbol;

output_row.Description = input_row.Description;

output_row.Rating = input_row.Rating;

output_row.Site = input_row.Site;

output_row.OwnerId = input_row.OwnerId;

output_row.CreatedDate = input_row.CreatedDate;

output_row.CreatedById = input_row.CreatedById;

output_row.LastModifiedDate = input_row.LastModifiedDate;

output_row.LastModifiedById = input_row.LastModifiedById;

output_row.SystemModstamp = input_row.SystemModstamp;

output_row.LastActivityDate = input_row.LastActivityDate;

output_row.CustomerPriority__c = input_row.CustomerPriority__c;

output_row.SLA__c = input_row.SLA__c;

output_row.Active__c = input_row.Active__c;

output_row.NumberofLocations__c = input_row.NumberofLocations__c;

output_row.UpsellOpportunity__c = input_row.UpsellOpportunity__c;

output_row.SLASerialNumber__c = input_row.SLASerialNumber__c;

output_row.SLAExpirationDate__c = input_row.SLAExpirationDate__c;

output_row.Location__Latitude__s = input_row.Location__Latitude__s;

output_row.Location__Longitude__s = input_row.Location__Longitude__s;

//CREATE PDF CODE - FROM ITEXT PDF EXAMPLE        

        ByteArrayOutputStream bos = new ByteArrayOutputStream();

        Document document = new Document();

        // step 2

        PdfWriter.getInstance(document, bos);

        // step 3

        document.open();

        // step 4

        document.add(new Paragraph("Account Name: " + output_row.Name + "/nAccountID: " + output_row.Id));

        // step 5

        document.close();

        

        byte[] bytes = bos.toByteArray();

        String stringToStore = new String(Base64.encodeBase64String(bytes));

        output_row.Content_Body = stringToStore;

        output_row.FileName = input_row.Name + ".pdf";

Step 5 - Insert the PDF into the Attachment Record 

Finally we pass along the values using a tMap component and the tSalesforceOutput component for Attachments. We map the output row fields from the tJavaRow onto our tSalesforceOutput row.

tMap Component

The complete demo job looks like this:

Step 6 - Export the Job and Run

The last step is exporting the job and running it on your machine. If you export the job design as a zip, you can get to the AccountPDFJob.sh file which executes the job. This video shows it running and you can see the final generate PDF attachment in SFDC.

I hope this helps folks. There are a lot of uses for this. You could generate PDF documents out of SFDC for all sorts of things with this approach. You could build a job that merges your SFDC data with an ERP and create a PDF invoice for example. Or you could use the Chatter objects and post files to a persons Chatter feed. There are lots of scenarios.