What the function basically does (with some more detail), it goes through each workarea and exports certain data from all tables that is used in the reporting engine. I would need to compare what percentage is the collection of the data, and how much of it is the creating of the XML file.

At first I was using memory allocation (lcValue = lcValue + lcAddString. But of course that's the worse as soon as the string is growing, so I changed it using FPuts.

To optimize it, I create the sections in memory and then add the section to the file. So it does go pretty fast and most importantly, does not slow down as the size grows.


>You probably use one of the faster ways to build large results (see my reply to R. Strahl for 2 typical ways), but it still might be a good idea to check if your code cannot be speeded up with tools inherent to the runtime. If you cannot post the code, perhaps by describing the base working and when/how slowdowns occur we can see something overlooked.
>
>regards
>thomas
>
>>Would it be an idea to first implement the class with the typical read/write functions and do the string manipulations later? For my purpose I would need the read/write to 95%, while the others are going to be exceptions. That would also give a proof of concept early on, so you can be sure you are on the right track.
>>
>>I would like to compare the writer with the code I am using now to create a fairly large XML file. I am curious what the actual difference will be.