Do you get the same (bogus) results when you do the same thing by hand within Excel? Do you get the same (bogus) results if you record a macro in Excel to do the same thing?

Generally recording an Excel macro to perform a task, then modeling your Perl script on the macro produces best results.

I have seen this type of annoying behaviour in both Excel and Word which leads me to think that it has something to do with OLE itself. I have several programs where I had to put in a sleep 3 here and a sleep 5 there, which sucks of course. If you enable the visible property of the application thing gets even worse.

From what I've gathered perl's interface to OLE is "single-threaded", aka methods shouldn't return until done, but that is not the case obviously.

There are some properties in Word/Excel which enables "background" processing, mainly printing and spell checking, that you could try to disable, this is just a hunch on my part that may improve your situation or most probably not.

I have tried extensively to find some property that would indicate if the OLE engine is in idle state which would be pollable, mainly as a workaround, but have come up with nothing so far.

I am interested to hear what progress you make, collecting evidence to crack this issue I guess.

Actually the problem I found out was that it was not updating the charts after opening the file through the script as it normally do when you open the file by double clicking. So, I write down a macro to update the charts and then called that macro in the script and the problem is solved now.