I wonder if your date formatting could be made a bit easier, and minimize using string functions like left() and len().
Will this work instead?
Date(Date#([date/time], 'MMM DD, YYYY hh:mm:ss tt PDT'))
One thing to keep in mind int his scenario is that you are loading timestamp values. These are numerically stored decimal values. Date are stored as integers.
My current timestamp looks like 19 May 2015 13:17:41, and has a underlying numerical value of 42143.553946759.
=Num(Date#('19 May 2015 13:17:41', 'DD MMM YYYY hh:mm:ss'))
If I have today's date 19 May 2015, the numerical value is 42143.
=Num(Date#('19 May 2015', 'DD MMM YYYY'))
You should store date values as proper dates, meaning that they have no time associated with them. Use the Floor() function to eliminate the decimal part of the value. And for clarity perhaps use the Timestamp#() function to load the first step.
Date(Floor(Timestamp#([date/time], 'MMM DD, YYYY hh:mm:ss tt PDT')))
If you then need the time and/or timestamp store them in separate fields. To make the timestamp into a time, you need to make the integer part a 0, which can be done with the Frac() function.
Timestamp(Timestamp#([date/time], 'MMM DD, YYYY hh:mm:ss tt PDT'))
Time(Frac(Timestamp#([date/time], 'MMM DD, YYYY hh:mm:ss tt PDT')))
The problem with combining your charts is that you use different dimensions in the charts. The values are not aggregated over the same timeline in this case, they look the same but are different. I do not have any quick solution to this.
I probably would have tried to adapt the data model to allow for the comparison to make both cases run towards the same calendar. This is how ever potential quite complicated, and would require analysis of the application file and data structure.
To enable yourself on more advanced data modeling I would suggest looking it attending the "Data Modeling for Qlik Sense" training at Qlik Training