Solution suggestion ofor problem with format_csv() method in orion/data.py

[jc4000]

Line 113 data['timestamp'] = df[timestamp_column_name].astype('int64').values

throws error if the df[timestamp_column_name] is in YYYY-MM-DD HH:MM:SS format. The typecast to int(64) doesn't work unless df[timestamp_column_name] the method pd.to_datetime() need to be used.

adding the following code chunk before line 113 resolves the problem.
if pd.to_datetime(df[timestamp_column_name], format='%Y-%m-%d %H:%M:%S', errors='coerce').notnull().all():
df[timestamp_column_name] = pd.to_datetime(df[timestamp_column_name], format='%Y-%m-%d %H:%M:%S')