Can the organizers explain the train and test dataset much better please?

I find the train dataset has data from 1st Aug 2016 to 31st July 2018, whereas the test dataset is from 1st Aug 2016 to 31st March 2018. Why is the test dataset in the past compared to the train dataset?

But considering from example one store and one day in both train and test, none of the time period in the column 'time' match in both train and test. What does this mean?

What is the period for which the predictions are required? It's not clear when I read it as 100 days in the evaluation.

Please provide a clear explanation. Thanks