Oct-20-2020, 08:17 AM
(This post was last modified: Oct-20-2020, 08:17 AM by AlekseyPython.)
Python 3.8.5, numpy 1.19.2, numba 0.51.2
I use method from official manual:
I use method from official manual:
import numpy as np import numba as nb #get data from txt- file regular_expression = _create_regular_expression(ticker, by_mask_ticker) dtype = _create_dtype() array_data = np.fromregex(file=path, regexp=regular_expression, dtype=dtype) #add new column lenght = array_data.shape[0] array_for_date_time = np.empty(lenght, dtype='datetime64[s]') array_data = recfunctions.append_fields(array_data, 'date_time', array_for_date_time) #fill new column fill_date_time_column(array_data) @nb.jit(nopython=True) def fill_date_time_column(array_data): for current in array_data: s = current['YEAR'] + '-' + current['MONTH'] + '-' + current['DAY'] + 'T' + current['TIME'] #d = np.datetime64(s) #ERROR: Cannot cast unicode_type to datetime64[] #d = nb.types.NPDatetime(s) #ERROR: Unknown attribute 'NPDatetime' of type Module(<module 'numba.core.types' #d = nb.types.NPDatetime(Y=current['YEAR'], M=current['MONTH'], D=current['DAY']) #ERROR: Unknown attribute 'NPDatetime' of type Module(<module 'numba.core.types' d = nb.NPDatetime(s) #ERROR: module 'numba' has no attribute 'NPDatetime' current['date_time'] = dIf I read function like this (without numba):
def fill_date_time_column(array_data): for current in initial_data: s = current['YEAR'] + '-' + current['MONTH'] + '-' + current['DAY'] + 'T' + current['TIME'] d = np.datetime64(s) current['date_time'] = d, then all work ok. How I can fill column by datetime values?