Calculate total Customer Count from accumulate in and out counts at fixed time interval

[#1]

Hi, Community,

i have the followng sample data...

i would like to calculate the hourly , daily and monthly total customer count  with the below data,

i have create a cyclic timegroup with month, day and hour dimension, but i dont know how to write the expression,...

Site	raw_tl_history_datetime	tl_history_datetimeforcustomercount	tl_history_counter_in	tl_history_counter_out
A	2016-09-05 09:30	5/9/2016	0	0
A	2016-09-05 10:00	5/9/2016	5	3
A	2016-09-05 10:30	5/9/2016	6	5
A	2016-09-05 11:00	5/9/2016	14	14
A	2016-09-05 11:30	5/9/2016	24	23
A	2016-09-05 12:00	5/9/2016	34	25
A	2016-09-05 12:30	5/9/2016	44	36
A	2016-09-05 13:00	5/9/2016	46	40
A	2016-09-05 13:30	5/9/2016	48	46
A	2016-09-05 14:00	5/9/2016	49	48
A	2016-09-05 14:30	5/9/2016	53	49
A	2016-09-05 15:00	5/9/2016	55	54
A	2016-09-05 15:30	5/9/2016	58	56
A	2016-09-05 16:00	5/9/2016	62	57
A	2016-09-05 16:30	5/9/2016	65	59
A	2016-09-05 17:00	5/9/2016	67	60
A	2016-09-05 17:30	5/9/2016	70	60
A	2016-09-05 18:00	5/9/2016	76	61
A	2016-09-05 18:30	5/9/2016	80	62
A	2016-09-05 19:00	5/9/2016	88	63
A	2016-09-05 19:30	5/9/2016	90	64
A	2016-09-05 20:00	5/9/2016	111	67
A	2016-09-05 20:30	5/9/2016	121	100
A	2016-09-05 21:00	5/9/2016	131	111
A	2016-09-05 21:30	5/9/2016	144	144
A	2016-09-06 09:30	6/9/2016	0	0
A	2016-09-06 10:00	6/9/2016	6	3
A	2016-09-06 10:30	6/9/2016	8	5
A	2016-09-06 11:00	6/9/2016	16	14
A	2016-09-06 11:30	6/9/2016	20	18
A	2016-09-06 12:00	6/9/2016	35	32
A	2016-09-06 12:30	6/9/2016	45	44
A	2016-09-06 13:00	6/9/2016	50	46
A	2016-09-06 13:30	6/9/2016	51	47
A	2016-09-06 14:00	6/9/2016	53	50
A	2016-09-06 14:30	6/9/2016	57	51
A	2016-09-06 15:00	6/9/2016	59	52
A	2016-09-06 15:30	6/9/2016	60	53
A	2016-09-06 16:00	6/9/2016	78	67
A	2016-09-06 16:30	6/9/2016	80	68
A	2016-09-06 17:00	6/9/2016	88	70
A	2016-09-06 17:30	6/9/2016	90	80
A	2016-09-06 18:00	6/9/2016	92	88
A	2016-09-06 18:30	6/9/2016	94	89
A	2016-09-06 19:00	6/9/2016	99	91
A	2016-09-06 19:30	6/9/2016	104	93
A	2016-09-06 20:00	6/9/2016	111	99
A	2016-09-06 20:30	6/9/2016	131	120
A	2016-09-06 21:00	6/9/2016	140	133
A	2016-09-06 21:30	6/9/2016	145	145

for example, at the 10 hour data point, it should show 19( last record of the hour( indicate by the first record of the next hour)-first record of the hour)..

calculate from :

14-5=9 +

16-6=10

at the 5 day data point, it should show 144..

144-0 =144( last record of the day - first record of the day)

is that possible...

[sunny_talwar]

So, based on the dimension you are on, you would want to see the This Period - Last Period?

[#3]

thanks for replying sunny.

sunny , i am thinking to use last period -this period, because for example, when i change to hour dimension of the cyclic group, it shows the total number of customer walked in the store at that hour period, but the record is taken in 30 min interval...so.i think..e.g at the 10 hour data point, use the last period of the record(since no record is taken at 10:59, so use the first record of the next hour, that is  the record taken at 11:00)-first record of the hour(the one record taken at 10:00)..

the "in" column is the accumulate customer walked in, so when i find the difference between this 2 time , i can know how many customers have been walked in the site at the 10th hour period, that is from 10:00 to 10:59...

i have tried for a while but cannot fix it...

[#4]

i guess it also make sense  to use max(this period) -min(this period) and us a line chart to present it, like attachement testcount3.qvw...Thanks~