lost one second when parse datetime

[ohmycloud]

save the ts column into test.xlsx

ts
2023-10-31 00:39:00

convert test.xlsx to test.csv:

xlsx2csv test.xlsx test.csv  --dateformat '%Y-%m-%d %H:%M:%S'

the datetime 2023-10-31 00:39:00 is convert to 2023-10-31 00:39:00, with one second lost(run head test.csv):

ts
2023-10-31 00:38:59

[weeinn]

I encountered the same problem

before convert, use pandas read excel

import pandas as pd
file_path = "C:\\Users\\acodingcat\\Desktop\\test.xlsx"
df = pd.read_excel(file_path)

current dataframe

after convert, use xlsx2csv convert xlsx to csv file

from xlsx2csv import Xlsx2csv
temp_path = "C:\\Users\\acodingcat\\Desktop\\test.csv"
Xlsx2csv(file_path,
         dateformat="%Y-%m-%d %H:%M:%S",
         skip_hidden_rows=False,
         outputencoding="utf-8").convert(temp_path, sheetname="Sheet1")
df = pd.read_csv(temp_path)

after convert, some data lost one second accurary

[lexxxel]

I encounter the same issue:
the 'raw string' 11.04.2024 08:40 gets converted via the Value() function in another cell. Libreoffices shoes the correct solution, but xlsx2csv gives me: 11/04/2024 23:39

if I use the dateformat: 11.04.2024 23:39:59. It does not happen with every date, but allways with xx:10, xx:40 and xx:00

[lexxxel]

I did some debugging and it is a rounding error:

so libreoffice does calculate 10 digits (in my example 8 ones at the end). xlsx2csv then caluclates the date in line 876 (version 0.8.4). However, if I enter 11 digits, by adding another 1, the result would be correct