How to perform Arithmetic operations with Operators in the dataset.

[shasank]

Hi Community, 

I am trying to work on adding a datetime variable to adjust for a time zone. 

 DateTime	        Zone 
31OCT19:18:08:00	+02
05NOV19:06:00:00	+02
05NOV19:12:15:00	-05
08NOV19:20:30:00	+02
13NOV19:12:52:00	+08
15NOV19:17:44:00	+02
18NOV19:14:33:00	-05
20NOV19:06:21:00	+02
20NOV19:14:22:00	+08
22NOV19:14:25:00	+02

The issue is that the observations already have the arithmetic operators in the values. I need help in adding the Zone variable(Hours) to the Datetime variable. 

Any help is appreciated. 

Thank you. 

[PeterClemmensen]

is zone numeric or character?

[shasank]

It is in a character format.

[novinosrin]

data have;
input  DateTime :datetime20.	        Zone $;
format datetime datetime20.;
cards; 
31OCT19:18:08:00	+02
05NOV19:06:00:00	+02
05NOV19:12:15:00	-05
08NOV19:20:30:00	+02
13NOV19:12:52:00	+08
15NOV19:17:44:00	+02
18NOV19:14:33:00	-05
20NOV19:06:21:00	+02
20NOV19:14:22:00	+08
22NOV19:14:25:00	+02
;

data want;
 set have;
 k=sign(input(zone,8.));
 New_DateTime=sum(datetime,k,hms(abs(k),0,0));
 format New_datetime datetime20.;
 drop k;
run;

 Hi @shasank  I overlooked the data by mistake and so I have edited

[shasank]

data want;
 set test;
 length Zone $3.;

	Zone = substr(DF_datetime,max(1,length(DF_datetime)-2));

	k = sign(input(Zone,8.));
 DF_deldttm =sum(DF_deldttm,a,hms(abs(k),0,0));

 format DF_deldttm  datetime.;
		 
run;

Output: 
Date/time of delivery, DF	Zone	a	Datetime
31OCT19:18:08:00	2	1	31OCT19:19:08:01
05NOV19:06:00:00	2	1	05NOV19:07:00:01
05NOV19:12:15:00	-5	-1	05NOV19:13:14:59
08NOV19:20:30:00	2	1	08NOV19:21:30:01
13NOV19:12:52:00	8	1	13NOV19:13:52:01
15NOV19:17:44:00	2	1	15NOV19:18:44:01
18NOV19:14:33:00	-5	-1	18NOV19:15:32:59
20NOV19:06:21:00	2	1	20NOV19:07:21:01
20NOV19:14:22:00	8	1	20NOV19:15:22:01
22NOV19:14:25:00	2	1	22NOV19:15:25:01
The zone variable is going off. 

[novinosrin]

Take the solution given by @Tom  or @ballardw 

[ballardw · Accepted solution]

Please show what you expect the value results to be.

I might guess that you want something similar to this assuming your current datetime variable is actually a SAS datetime value, numeric with a datetime format assigned.

data have;
   input DateTime datetime.  Zone $;
   format datetime datetime16.;
datalines;
31OCT19:18:08:00	+02
05NOV19:06:00:00	+02
05NOV19:12:15:00	-05
08NOV19:20:30:00	+02
13NOV19:12:52:00	+08
15NOV19:17:44:00	+02
18NOV19:14:33:00	-05
20NOV19:06:21:00	+02
20NOV19:14:22:00	+08
22NOV19:14:25:00	+02
;

data want;
   set have;
   datetime = intnx('hour',datetime,input(zone,best.),'s');
run;

The INTNX function increments time, date or datetime values by a given interval, 'hour' in this case. The Input turns the zone into a numeric value assuming hours, and the 'S' parameter means "same" for keeping minutes.

[Peter_C]

Have a look at how SAS supports adjusting for timezones at documentation.sas.com .... programming in sas9.4
https://documentation.sas.com/?cdcId=pgmsascdc&cdcVersion=9.4_3.5&docsetId=nlsref&docsetTarget=p12pv...

[Tom]

The normal numeric informat will convert those strings into numbers just fine.  So this will create a numeric variable OFFSET with the number of hours (positive or negative).

offset=input(zone,32.);

You could then use INTNX() function or simple arithmetic to add the offset value to your current datetime value.

universal_dt = intnx('hour',datetime,offset,'s');
universal_dt = datetime + offset*'01:00't ;