I wonder if someone could take a minute out of their day to give their two cents on my problem.
I would like some suggestions on what would be the best data structure for representing, on disk, a large data set of time series data. The main priority is speed of insertion, with other priorities in decreasing order; speed of retrieval, size on disk, size in memory, speed of removal.
I have seen that B+ trees are often used in database because of their fast search times, but how about for fast insertion times? Is a linked list really the way to go?
You might want to look into HDF5 (Hierarchical Data Format). It's well suited for Time Series data. Implementation wise, it uses B Trees.