Performance about distributed database on a single physical machine?

Question

Currently I\'m working with time-series data(Tickers) and stored the pandas dataframe in .h5 and .parquet format. My workload is basically querying the