Agreed, but you know the route I'd go? I'd put my check boxes in a container and then put the container in a grid column...I don't know why, but I have much better luck with that.

I have one really painfully complicated form with pgframes with a grid that has one column with a container. The container has 30 (?) controls, including checkboxes, cbos (this was before I knew about all the problems with cbos), an edit box and even another grid. It was awkward in 3.0, but in 5.0 it's actually darned sweet given the situation.

>This definitely makes things easier, I can imagine. However, if the grid only has a few fields on it, it's just more user-friendly, IMO, to allow editing directly on the grid. For example, in my current project, the main user action on the gridded table will be to check checkboxes. It's so much easier for the user just to be able to check the boxes on the grid than check the right grid row, then go up to the input area.

Hm. I use the tiny little column :-) Laziness mostly. Is your button in a column of the grid? As for suggestions, I haven't done this, so I can't say offhand. I'll think about, but I bet one of the other folks here has done something similar and can advise :-)

>One problem I had today was deleting records from the grid. I won't use the built-in method of clicking on a tiny little column, so I have a Delete button. My problem is determining whether a grid row has been selected, i.e. has a mark in the leftmost column. I've resorted to setting a record number property in the grid's AfterRowColChange (when a row has been selected), and having the form clear that property when another record is accessed (ugly!!). Do you have a better suggestion? In particular, how do you decide what to fill in your child input area with when a parent record is first displayed?