Recently, I had an idea to quantify OKEX, the ETH-quarter futures, going up 21 years, up 20 years, down 19 years, and then look carefully at the trading records, and find that OKEX's 1min data for 19 years jumped a lot, so I changed the position of HuobiDM ETH-quarter, 19 years in the forecast strategic gains are normal, 20 years again jump.
In summary: the 19th-quarter okex, 1min data and a lot of jumps, 18 years unknown, tokens, 20 years jumps.
This is just 1 minute of data. So the question arises, why do you need to split the data of the exchange instead of collecting more coins? Suppose the ETH quarterly data, why not carefully modify a complete one, but divide it into so many exchanges? At this point, because your data problem excludes an inspiration I think it's not worth it, I strongly recommend that you repair a complete data, without having to split so many exchange reviews at all, now it's a bit confusing, which data is complete, reviews need to re-test both OKEX and HUOBI, and it's a bit more relaxed.
The grassOkayx's historical data collection is not good, it does need to be organized.