Open jingedawang opened 1 year ago
There is a problem with qlib data. If a stock was suspended for some days, the features of this stock in these days are NaN. But if the stock is still suspended till today, the feature will be empty. I guess when the stock is resumed, the empty feature will be filled with NaN.
So, the solution in fix_missing_data
will focus on the NaN cases, this is used to recognize which period should be fixed.
Running this command line
python ~/projects/qlib/scripts/data_collector/yahoo/collector.py update_data_to_bin --qlib_data_1d_dir ~/.qlib/qlib_data/cn_data --trading_date 2022-09-20 --end_date 2022-09-21
didn't fix the data missing problem. The next run still missing this data. Not sure why. Let's keep an eye on it later.
If due to some reason, some days' trading data is missing, we need to re-download them in a fixed time schedule.