Suppose, I have large (about 1 GB) strings in ASCII encoding. I have to choose whether to store them as separate blob fields or split each one of them into smaller pieces and then store them as several blob fields. After I get them in my DB, I will have to update these strings (not completely, just some small parts) and select some parts of them. It seems to me that standard MySQL functions REPLACE() and SUBSTRING() should fit just well for these purposes. What I want to ask is does a blob field get loaded completely into the memory of server when these functions are invoked? For me it's important because if my database is accessed by several users via a client application and they modify different strings simultaneously, the server may run out of memory. And if it's true I definitely have to divide my strings into parts. Also it will lead to more sophisticated queries on the client side.
Surely, if you come up with a different solution for the problem of storing large strings I'd really appreciate it.