converting character $4 to military time

[JamDOT]

I have tried the following in Windows 9.3:

data want;
set have;
Varnew=input(Varold,time.);
FORMAT Varnew time5.;
run;

The procedure runs, but only one observation and two variables result when there should be 1000s.

NOTE: Numeric values have been converted to character values at the places given by:
(Line):(Column).
548:16
NOTE: There were 1 observations read from the data set have.
NOTE: The data set want has 1 observations and 2 variables.
NOTE: DATA statement used (Total process time):
real time 0.03 seconds
cpu time 0.04 seconds

The data is 4 characters long and in military time. I ultimately want to convert the observations to numeric format so I can group them by the hour. They are currently by the minute.

[RW9]

I don't know what you mean by military time?  This is where showing example data helps.  

[JamDOT]

The character data is 4 characters long:

0001

0002

0003

....

2356

2357

2358

2359

I want to convert this character variable to a numeric time format.

[Shmuel]

From your log lines

NOTE: There were 1 observations read from the data set Varold.

NOTE: The data set Varnew has 1 observations and 2 variables.

it seems that you have not psted the right code.

According to posted code varold and varnew are variables not datasets !

[JamDOT]

you are correct:

NOTE: Numeric values have been converted to character values at the places given by:

(Line):(Column).

548:16

NOTE: There were 1 observations read from the data set have.

NOTE: The data set want has 1 observations and 2 variables.

NOTE: DATA statement used (Total process time):

real time 0.03 seconds

cpu time 0.04 seconds

[Rick_SAS]

Your log messages do not match the program that you submitted.

The TIME. format displays time in 'military time' (24-hour clock) but does not print a leading zero for 06:00.

If you want the leading zero, use the TOD. format.

I suggest you use the ANYDTTME. informat to convert the strings to a time value, as follows:

data t;
length s $8.;
input s;                 /* a string */
t = input(s, ANYDTTME.); /* a (numeric) time value */
format t tod8.;          /* tell SAS to format the number as a 24-hour time */
datalines;
23:02:03
22:57:16
06:00:00
04:40:22
run;

proc print;run;

[JamDOT]

why do datalines need to be included? Isn't the data being fed from 't'?

[Rick_SAS]

I used DATALINES so that everyone could see the data.  If you have a data set 'have' that contains the variable s, you can use

data t;
SET have; /* the 'have' dataset contains the character variable s */
t = input(s, ANYDTTME.); /* a (numeric) time value */
format t tod8.; /* tell SAS to format the number as a 24-hour time */
run;

[JamDOT]

605 data want;

606 set have;

607 Varnew=input(varold,anydttme.);

608 FORMAT varnew tod8.;

609 run;

NOTE: There were 1167321 observations read from the data set have.

NOTE: The data set new has 1167321 observations and 113 variables.

NOTE: DATA statement used (Total process time):

real time 3.88 seconds

cpu time 3.52 seconds

　

610 proc freq data= want;

611 tables varnew/nocum;

612 run;

NOTE: There were 1167321 observations read from the data set WORK.TRAVELTIME.

NOTE: PROCEDURE FREQ used (Total process time):

real time 1.24 seconds

cpu time 1.21 seconds

The FREQ Procedure
endtimeN Frequency Percent 
      [][][] /*3 column table here with blank cells*/
Frequency Missing = 1167321 

  The output table still produces no data, BUT at least all observations are now being read.

[Rick_SAS]

If the data are really of the form 

0001

0002

etc

then those are invalid data strings, which is why the informat is converting them to missing. Try creating a new string that inserts a colon between the second and third characters, like this:

data t;
length varOld $4. s $5.;
input varOld;                 /* a string */
s = catx(":", substr(varOld,1,2), substr(varOld,3,2));
t = input(s, ANYDTTME.); /* a (numeric) time value */
format t tod8.;          /* tell SAS to format the number as a 24-hour time */
datalines;
0001
0002
0003
2356
2357
2358
2359
;

proc print; run;

[JamDOT]

it works with datalines but when I put the set statement in, I get errors again.

24   data want;
25   set have;
26   length varold $4. varnew $5.;
27   input varold;
28   varnew =catx(":", substr(varold,1,2),substr(varold,3,2));
29   want=input(varnew,anydttme.);
30   format want tod8.;
31

ERROR: No DATALINES or INFILE statement.
NOTE: The SAS System stopped processing this step because of errors.
WARNING: The data set want may be incomplete.  When this step was stopped there were
         0 observations and 114 variables.
WARNING: Data set want was not replaced because this step was stopped.
NOTE: DATA statement used (Total process time):
      real time           0.04 seconds
      cpu time            0.04 seconds

32   proc print;
33   run;

NOTE: There were 1 observations read from the data set want.
NOTE: PROCEDURE PRINT used (Total process time):
      real time           0.03 seconds
      cpu time            0.01 seconds

[BrunoMueller · Accepted solution]

You can use the HHMMSS. informat to read a hhmm char value and translate it to a SAS time value.

See code below:

data t;
  do charTime =
    "0001"
    ,"0002"
    ,"0003"
    ,"2356"
    ,"2357"
    ,"2358"
    ,"2359"
  ;
  sasTime = input(charTime, hhmmss4.);
  output;
 end;
 format sasTime tod8.;
run;

[JamDOT]

where in the code does the data file go if I am not using datalines (very much a novice here)