Hi Ravi,
I can think of one way to do it but it requires mapping for it to work. Setup the source interface with two mappings - one for the target header and one for the target data rows - add an interface split and set the option for maintaining order at runtime. In this way you can ensure that you have one header target that will run first and then for however many data splits you have you will keep appending that data to the same file.
Regards,
Ryan Crosby