# Is it necessary to sort time series data on datetime stamps?

Say I have a time-series data frame for a bank that records number of cases filed for 16 districts as below (cases can be for many purposes - loans, credit cards, real estate, etc.):

application_date    branch_id    district    case_count
2017-04-01              1        A           40.0
2017-04-03              1        A           5.0
2017-04-04              1        A           4.0
2017-04-05              1        A           113.0
2017-04-07              1        A           76.0


But if I sort the above data by application_date, like so:

application_date    branch_id    district    case_count
2017-04-01              1         A           40.0
2017-04-01              62        B           5.0
2017-04-01              63        C           4.0
2017-04-01              64        D           113.0
2017-04-01              65        E           76.0


Considering that branch_id and district are categorical variables, does sorting affect performance of standard ML models like random Forests and gradient boosted models as well as their predictive accuracy?