turn char date into SAS date format and calculate time interval for each customer

[ccnky123]

The original table is like:

cst_id	date
A01	2015/5/6 9:13:59
A02	2015/5/7 10:09:8
A01	2015/5/7 15:12:02
A02	2015/5/9 15:09:45
A01	2015/5/12 23:07:33
A02	2015/5/13 14:03:16

and finally get:(the real table for column date may be disorder)

cst_id	date	date1	interval
A01	2015/5/6 9:13:59	06may2015 9:13:59	.
A01	2015/5/7 15:12:02	07may2015 15:12:02	107863
A01	2015/5/12 23:07:33	12may2015 23:07:33	460531
A02	2015/5/7 10:09:01	07may2015  10:09:01	.
A02	2015/5/9 15:09:23	09may2015 15:09:23	190837
A02	2015/5/13 14:03:16	13may2015 14:03:16	341611

Data step will be preferable! thank you for your attention 

[VDD]

proc sort data=have;

by cst_id date;

run;

now do the calculations using lag.

[Kurt_Bremser]

---

@ccnky123 wrote:

The original table is like:

cst_id	date
A01	2015/5/6 9:13:59
A02	2015/5/7 10:09:8
A01	2015/5/7 15:12:02
A02	2015/5/9 15:09:45
A01	2015/5/12 23:07:33
A02	2015/5/13 14:03:16

and finally get:(the real table for column date may be disorder)

cst_id	date	date1	interval
A01	2015/5/6 9:13:59	06may2015 9:13:59	.
A01	2015/5/7 15:12:02	07may2015 15:12:02	107863
A01	2015/5/12 23:07:33	12may2015 23:07:33	460531
A02	2015/5/7 10:09:01	07may2015  10:09:01	.
A02	2015/5/9 15:09:23	09may2015 15:09:23	190837
A02	2015/5/13 14:03:16	13may2015 14:03:16	341611

Data step will be preferable! thank you for your attention 

---

Use the scan() fubction to dissect yor "date" string. input() the first "word" with yymmdd10., and the second with time8.

Then build the datetime value with dhms(date,0,0,time).

After that, use lag() as advised.

For code, provide your example data in usable form (data step with datalines).