I overwrote this data importing a high quality tick data from Dukascopy for purpose of backtesting.
Now I have two distinct datasets: one is Market Replay set, stored only in a market replay data, another one is a historical data, stored in NT data base.
I cannot overemphasize how faulty is this situation. Basically I see 3 options to treat this situation:
1) Historical data import will include creation of Market Replay data files according to the imported data. It shouldn't be hard.
2) Market Replay data will be purged.
3) Leave the existing situation, but warn a user before importing a historical data about two distinct data sets going to be created.
I prefer the 1-st.
Of course an idiot proof solution is not possible. I can import a historical data and copy another market replay data set into its directory, but then I'll be a fully responsible idiot user.

Comment