すべての注文
ルール
指定
I've got convert_csv_to_mt.py script (in Python 3) which incorrectly converting the CSV files into FXT. Basically all timeframes have the same size, but they shouldn't. The size should be smaller as the timeframe increases.
Currently FXT files for one year have 80MB for all the timeframes which is not correct.
Notes: This was working some time ago. You may compare the logic with similar scripts.
Reproducible steps using these commands (given the linked scripts and downloaded into the folder):
1. Download CVS files for 1 month:
python3 dl_bt_dukascopy.py -c -p EURUSD -y 2014 -m 1
2. Combine all CSV files into one file (Linux, on Windows you can just try with one file):
find . -name "*.csv" -print0 | sort -z | xargs -0 cat > all.csv
3. Convert CSV into FXT (should be less than 2 mins):
time python3 convert_csv_to_mt.py -v -i all.csv -s EURUSD -t M1,M5,M15,M30,H1,H4,D1,W1,MN -f fxt4
4. Now all the files have 77MB, but the size should decrease as the timeframe is higher.
Requirements:
- the script should not take higher time of the conversion process than the current (since it has less processing)
- the output format should be valid and pass the validation tests (which it is currently),
- the output format should work fine under MT4 platform (as it is currently)