first
i am sorry guys to answer so late ,thanks for your reactions
the original reason was that different programs were generating and using different file extensions and i have to adapt for them when using data from or feed data to these projects .(suppliers specification)
in the meantime however i standarized the input( and output) for my own project on .txt and simply adapt the name of te file when needed.
ps i want to split files because the are too big
the reason was that up till now i have used stringlist to open big files (> 70mb) ,and had put data in arrays ( data are 30 -50 parameters on a one minute base over more then 20 years resulting in too big arrays and very slow and memory problems.
I was forced to split the files by hand
It may be that using Tfilestream for picking data (used for statistical calculations)is the better option here .
Some time ago i standarized the storage (size per day) , but still using Tstringlist.
So splitting the file in smaller parts ,the same idea as Tfilestream is using buffers.
frederic