topic Re: Time variable into categorical in SAS Programming

Time variable into categorical

Rodrigue — Mon, 21 Jun 2021 12:35:28 GMT

Hi ! Please I need you help+++

I have a time variable HH:MM called "timestop" and I would like to create a new categorical variable "timestopnight" stating 1 if the timestop was between 23:00 and 7:00 and 0 if the timestop was between 07:00 and 23:00.

I have tried several codes like the following but none worked.

data dataset ; set dataset;

if timestop < '07:00't then timestopnight=1;

if timestop > '23:00't then timestopnight=1;

if '07:00't =<timestop =< '23:00't then timestopnight=0;

run;

Thanks a lot for your help 🙏

Rodrigue

Re: Time variable into categorical

tarheel13 — Mon, 21 Jun 2021 12:38:35 GMT

I think you need to use if else if.

Re: Time variable into categorical

tarheel13 — Mon, 21 Jun 2021 12:40:09 GMT

Can also do a boolean.

timestopnight=07:00't le timestop le 23:00't;

This will assign 1 if the condition is true and 0 if it's false.

Re: Time variable into categorical

Kurt_Bremser — Mon, 21 Jun 2021 13:19:05 GMT

Please post the complete log from that step.

If timestop is a SAS time, this will work:

timestopnight = ifn('07:00't le timestop le '23:00't,0,1);

Be observant if timestop is a datetime instead of just time.

Re: Time variable into categorical

Rodrigue — Mon, 21 Jun 2021 13:43:51 GMT

Thank you very much Kurt. It seems to work perfectly well for subjects who have a time in the timestop variable.

The only problem is that I have a 1 in the timestopnight column in patients who don't have value for the timestop variable.

I've tried to insert before but it didn't work.

timestopnight = .;

Could you help me with that ?

I have attached the log but the variable names are different since I've tried to simplify for the post.

Thanks again for your help+++

Rodrigue

Re: Time variable into categorical

Rodrigue — Mon, 21 Jun 2021 15:29:35 GMT

I found the solution adding

if not missing (timestop)

Re: Time variable into categorical

ballardw — Mon, 21 Jun 2021 15:54:26 GMT

MISSING values are always less than any actual value.

Always.