First, I want to say that the main MySQL tick data thread is here:
I don't want to have a split topic but I do think some discussion specific to using R and MySQL for tick data is applicable to the R subforum.
I've been storing tick data in MySQL for years but am recently expanding this project. Right now my database is around 50GB for 2 symbols (CL, ES) for 4 years of tick data.
I am going to expand this to around 20 symbols soon but wanted a better way of storing the data.
So I wrote this post:
Which I will go ahead and copy below for this thread:
I think that for storing light data, like daily bars, MySQL by itself is fine. But the issue is when you get into databases being 500GB - 1TB like I am looking at, better solutions are required.
Simply looking at column oriented databases is not the answer. I already went down that path (discussed in the main MySQL thread linked above). Besides, for tick databases you are almost always just doing straight select's of the entire range of data, say 1 day at a time, 1 week at a time, 1 month, 1 year whatever.
So I think what I am proposing makes more sense, but I have not had time to test it yet. Before I start heading that direction I wanted to hear from some others to see if what I've written makes sense or could be improved.