topic Re: How to write code to identify error in SAS Programming

How to write code to identify error

Bhavanaa44 — Sun, 27 May 2018 17:40:18 GMT

Data consists of subject Id by visit and date so how to identify wrong date by program. As after visit 1 is date of enrollment. And I have given only 1 subject details if there are more then how to identify discrepancy error.

Subjid sex age visit date
101 M. 25. 1. 30-05-2016
101. M. 25. 2. 29-06- 2016
101. M. 25. 3. 25-07-2016
101. M. 25. 4. 26-02-2016
101. M. 25. 5. 25-08-2016

Re: How to write code to identify error

Shmuel — Sun, 27 May 2018 18:30:15 GMT

I understand that the date is given in a char type variable, in a format of ddmmyy10.;

While converting the date into num type you ususlly use the ddmmyy10. informat, or any other informat.

You can code: num_date = input(char_date, <informat>, ??);

in case of date error you will get missing value.

check the num_date is it missing or not.

Re: How to write code to identify error

Cynthia_sas — Sun, 27 May 2018 18:31:06 GMT

Hi:
Some questions:
1) Why is there a period (.) after every value except the first row? There's no period after 101 in the first row. Is this intentional? A typo? Does it mean something?
2) How do you identify a discrepancy? Of the above 5 rows, where is the discrepancy and what it your logic?
3) You said "As after visit 1 is date of enrollment" -- that doesn't make sense to me -- there are 4 rows after visit 1 does that mean there are 4 dates of enrollment? I would have expected visit 1 to be the date of enrollment AND the date of the first visit and that for every other visit, the date just represents the visit date? I don't understand how you can define that "after visit 1 is date of enrollment".

What code have you tried? What is your desired output?

Cynthia

Re: How to write code to identify error

Kurt_Bremser — Mon, 28 May 2018 09:07:22 GMT

Converting the dataset to a data step and posting the resulting code here in the proper way (see my footnotes for both of these steps) will remove any ambiguity about your data and let us use it for testing. Additionally, post an example for how your datset should look like after processing.