When `process_only_new_candles` is enabled and no new candle is detected, `_analyze_ticker_internal` previously called `remove_entry_exit_signals` to clear signal columns. Since the dataframe is a fresh copy from `dp.ohlcv` (which only contains OHLCV data) and the return value is only used for basic validation (length, close price, date) in `analyze_pair`, this operation was redundant. Removing this call saves approximately 1.3ms per pair per iteration loop (when skipping analysis), avoiding unnecessary pandas column assignments. The test `test__analyze_ticker_internal_skip_analyze` was updated to reflect that signal columns are no longer added in this scenario. Consumers of analyzed data (e.g. `freqtradebot`) rely on `DataProvider` cache, which retains the previously analyzed (populated) dataframe. Co-authored-by: Corax-CoLAB <239841157+Corax-CoLAB@users.noreply.github.com>pull/12809/head
parent
9ed6a15464
commit
ea0cce77f7
Loading…
Reference in new issue