Why is the timedelta command generating an error on another computer?

I am currently working on two different computers, but one of the computers is running the same code on one of the two computers:

ValueError: cannot create timedelta string converter for [-1 days +23:05:00.000000000]

      

        
        
        
      

    

The error was caused by the command:

df_gt["Duration"] = pd.to_timedelta(df_gt[ "Duration"])

      

        
        
        
      

    

And df is created by importing csv via:

df_gt = pd.read_csv(gt_data, sep='\t', header=(0))

      

        
        
        
      

    

First, I thought the error was caused by a negative number. I've seen problems with negative numbers in the past, but it must have been the same problem on a different computer.

Then I wondered if this could be caused by another csv / time setting on this computer generating other formats, but when I print the value, I get exactly the same format (at least by visual inspection). I still think the problem is somewhere in this domain, but I'm not sure how another computer might affect this internal code.

Some information about the hardware:

The computer that everything runs on is running Windows 10 Home and is using Pycharm Professional 2017.1. The other is running Windows 7 Professional sp 1 and is using Spyder 2.3.1. The code is in Python 3.

Thanks in advance,

Jeff

EDIT:

The original question still had

.total_seconds()

      

        
        
        
      

    

attached to the team, it was by mistake. (Just an attempt by my self to resolve this)