The syntax of the command looks fine, I just tested it and the parsing of that time range comes out as expected.
24 hours is a really long time period for all flow data, what you might be seeing are the flows that have started from several days prior, but ended within the interval, but since you don't see past the first 10k the flows within the given time range never appear.
A simpler approach might be to choose much smaller time intervals, then optionally stitch the results together if you needed to see the complete view. But note this could be quite a time consuming process, especially if your NetProfiler is seeing very large numbers of flows.
Another approach would be to customize the script and add a higher "limit" argument to the run method. By default, as mentioned above this limit is set to 10000, but you can increase that number to get more flows out of the system.
Another thing to consider is that the flow start times in this report are irrespective of the time frame for the report. If, as in your example, you query from "7/07/16 0:00 to 7/07/16 23:59" and during that time there is an active flow that has been present on the network since "07/01/16 04:00:00 EDT' then you will see the UNIX time stamp for that time (1467345600.00) in the start_time column. If that is what you are seeing then this is expected. What would be problematic is a flow with a start and end time both prior to the query time specification.