Hey Pratiksha,
I'll try to clarify a bit further. Currently this is the process we are doing:
1. Getting parent rows
2. Getting each child row for each parent row, building a grid, and ripping the html and placing it into the table row of the first grid. Because we are placing regular html, the standard events that come from a grid are difficult to keep. Usually what I do is build the child grid with an edit hyperlink which populates a pop up window for edit, or redirects to a different approach. If we wanted to keep the features of the child grid, and have edit/cancel options, we would have to modify the code entirely. Pretty much instead of ripping the html we would first create dynamic events, then add the grid as a control to the parent grid row. This will allow proper viewstate and event registration. Next, on page load you would have to recreate the grid and childrows to synch with the viewstate. Then the automated events for edit and cancel would fire. It is much more complex, and I would recommend the first apprach.
I hope this helps. Sorry for the late reply.
Rajib