Solved: Reference DateTime for calculating occupancy - Page 2 - Qlik Community

dafnis14 · ‎2017-02-02

Hi All,

I would like to analyze occupancy percent of a parking lot .

My data include entrance and exit timestamps of cars.

I generated reference timestamps for each hour in the range.

My question is which is the best time resolution (i.e., hour, half an hour..) to analyze the occupancy?

This is due to the fact that the entry and exit times can vary, and I'm not sure how should I round them.

Please consider such cases:

Receipt_EntryDateTime	Receipt_ExitDateTime
04/03/2016 11:35:00	04/03/2016 12:33:00

Receipt_EntryDateTime	Receipt_ExitDateTime
06/04/2016 11:43:00	06/04/2016 12:27:00

Thanks!

sfatoux72 · ‎2017-02-03

Hi Marco,

I found just a little problem, but I found the solution too 😉

The problem

If you consider only one receiptKey (that start en end in the same day) with a BinSize 00:15 like that, the first bar is not correct :

The reason

It is due to the expression of the calculated dimension :

=Time(Floor(Time,vBinSize),'hh:mm')

The value used as a step by the floor function is not a finite value (00:15 ==> 0.010416666666666666666 ...), but since it is the finite value 0.010416666666667 that is actually used, there is in the end A slight offset that causes the problem found.

The solution

We need to work with minute in this expression :

=Time(Floor(Hour(Time)*60+Minute(Time), Hour(vBinSize)*60+Minute(vBinSize))/24/60,'hh:mm')

Like that the step is a finite value 15 and we have a correct result :

MarcoWedel · ‎2017-02-03

Hi,

yes, due to the limited resolution of the underlying numerical values, time arithmetics may show this behaviour.

Another solution might be:

=Time(Floor(Time,Time#(Time(vBinSize,'hh:mm'),'hh:mm')),'hh:mm')