Date: Fri, 11 Jan 2002 14:04:30 -0500
Reply-To: "Chakravarthy, Venky" <Venky.Chakravarthy@PFIZER.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: "Chakravarthy, Venky" <Venky.Chakravarthy@PFIZER.COM>
Subject: Re: Parsing a Date/Time Text Field
Content-Type: text/plain; charset="iso-8859-1"
Daniel,
In version 8.2, the DATETIME INFORMAT is capable of handling AM/PM.
113 data _null_;
114 x='May 21 2001 12:20:31:123AM' ;
115 y = input(substr(x,5,2)||substr(x,1,3)||substr(x,8,4)||"/"||
116 substr(x,13,8)||"."||substr(x,22,3)||" "||
117 substr(x,25,2),datetime40.) ;
118 date = datepart(y) ;
119 time = timepart(y) ;
120 put y= date= time= ;
121 run;
Y=1306023631.1 DATE=15116 TIME=1231.1229999
Kind Regards,
Venky
#****************************************#
# E-mail: venky.chakravarthy@pfizer.com #
# swovcc@hotmail.com #
# Phone: (734) 622-1963 #
#****************************************#
-----Original Message-----
From: Kitzmann, Daniel J. [mailto:kitzmann.daniel@MAYO.EDU]
Sent: Friday, January 11, 2002 11:49 AM
To: SAS-L@LISTSERV.UGA.EDU
Subject: Parsing a Date/Time Text Field
Dear Listers:
Suppose I have a dataset containing vars d1-d(n), where each var is a
date/time field stored as a text expression in the somewhat peculiar form:
Obs d1 d2
1 May 21 2001 12:00:00:000AM May 18 2001 11:01:24:126AM
2 May 21 2001 12:00:00:000AM May 18 2001 11:23:57:870AM
3 May 21 2001 12:00:00:000AM May 18 2001 4:29:14:216PM
4 May 21 2001 12:00:00:000AM May 11 2001 8:01:30:010AM
5 May 21 2001 12:00:00:000AM May 11 2001 8:14:01:020AM
Suppose, too, that for each var d(i), I want to parse the text string and
create two separate vars, u(i) and t(i), where u(i) is a SAS date var and
t(i) is a SAS time var. The following code achieves same:
265 data a;
266 set;
267 array d{2};
268 array e{2} $ 14;
269 array t{2};
270 array u{2};
271 do i = 1 to dim(d);
272 %let x = scan(d{i}, ;
273 e{i} = reverse(trim(&x 4)));
274 substr(e{i}, 3, 1) = ' '; ** sig. digit precision lost;
275 substr(e{i}, 6, 1) = '.'; ** at .000 s is trivial;
276 t{i} = input(reverse(e{i}), time14.);
277 u{i} = input((&x 2) || &x 1) || &x 3)), date9.);
278 end;
279 run;
NOTE: There were 5 observations read from the data set WORK.A.
NOTE: The data set WORK.A has 5 observations and 9 variables.
Obs d1 d2 e1
1 May 21 2001 12:00:00:000AM May 18 2001 11:01:24:126AM MA
00.00:00:21
2 May 21 2001 12:00:00:000AM May 18 2001 11:23:57:870AM MA
00.00:00:21
3 May 21 2001 12:00:00:000AM May 18 2001 4:29:14:216PM MA
00.00:00:21
4 May 21 2001 12:00:00:000AM May 11 2001 8:01:30:010AM MA
00.00:00:21
5 May 21 2001 12:00:00:000AM May 11 2001 8:14:01:020AM MA
00.00:00:21
Obs e2 t1 t2 u1 u2 i
1 MA 21.42:10:11 0 39684.12 15116 15113 3
2 MA 78.75:32:11 0 41037.87 15116 15113 3
3 MP 12.41:92:4 0 59354.21 15116 15113 3
4 MA 10.03:10:8 0 28890.01 15116 15106 3
5 MA 20.10:41:8 0 29641.02 15116 15106 3
My question is whether the foregoing can be done in a more efficient or
transparent way. I puzzled over trying to do the entire conversion in a
single pair of input functions (one each for date and time), with the
concomitant parsing functions nested inside as the input argument (as is
done with the u(i) var). But, as shown, I could not think of a way to do it
in a single line with the t(i) var.
Thank you kindly in advance for any solutions.
Cordially,
Daniel Kitzmann
kitzd@bigfoot.com
