adelorenzo/stable
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Reduce time taken for def backtest by 60%

Browse Source 
* on 2 weeks 5minute data

Performs bitwise filter on buy/sell rows in the df
before passing to the slower itertuples loop

the larger the dataset the greater the performance gain
This commit is contained in:
creslinux 2018-07-11 18:26:27 +00:00
parent bd46b4faf3
commit 29fdba7743
1 changed files with 15 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
freqtrade/optimize/backtesting.py
View File

@ -221,7 +221,21 @@ class Backtesting(object):
# Convert from Pandas to list for performance reasons # Convert from Pandas to list for performance reasons
# (Looping Pandas is slow.) # (Looping Pandas is slow.)
ticker = [x for x in ticker_data.itertuples()]
#ticker = [x for x in ticker_data.itertuples()]
"""
Use faster numpy to mask out rows we will not use in slower itertuples loop
Filter rows that are buy or sell to be processed
Also keep the last row from in any state for open trades
"""
mask = (ticker_data['buy'].values != 0) | (ticker_data['sell'].values != 0)
ticker_loop_data=ticker_data[mask]
end_row = DataFrame(ticker_data[-1:].values, columns=ticker_data.columns)
ticker_loop_data = ticker_loop_data.append(end_row)
ticker = [x for x in ticker_loop_data.itertuples()]
lock_pair_until = None lock_pair_until = None
for index, row in enumerate(ticker): for index, row in enumerate(ticker):