Compare multiple column date values and flag the earliest/smallest date variable

[kelly2016]

Hello SAS Community,
I'm a relatively new SAS programmer and I really need your help coding this!  I have a dataset that contains 5 columns, each representing a different disease diagnosis date in date format (date9.).  Some columns contain missing values.  I need to compare the diagnosis dates across these 5 columns when the value is not missing and flag or print out the earliest diagnosis date variable. Essentially, to find which disease diagnosis occurred first and flag that. 

Is there an easy way to code this?  See my sample data and expected output below. Thank you so much for your help!!

Data Example:

ID	Diag1	Diag2	Diag3	Diag4	Diag5
1	26-Oct-17		28-Oct-18	26-Mar-19
2	.	.	.	.	.
3	22-Dec-21	.	.	.	.
4	.	.	.	10-May-21	13-Nov-22
5	15-Oct-19	15-Oct-19	.	24-Jun-15	23-Oct-17
6	25-Dec-16	14-Oct-18	24-Nov-22	26-Dec-15	23-Oct-23

I would like the output to be as follows:

ID	Diag1	Diag2	Diag3	Diag4	Diag5	Flag
1	26-Oct-17		28-Oct-18	26-Mar-19		Diag1
2	.	.	.	.	.	Disease Free
3	22-Dec-21	.	.	.	.	Diag1
4	.	.	.	10-May-21	13-Nov-22	Diag4
5	15-Oct-19	15-Oct-19	.	24-Jun-15	23-Oct-17	Diag5
6	25-Dec-16	14-Oct-18	24-Nov-22	26-Dec-15	23-Oct-23	Diag4

[Reeza · Accepted solution]

data want;
set have;

length flag $40.;

array _diag(*) diag1-diag5;

if nmiss(of _diag(*)) ne dim(_diag) then do;
    *find smallest value;
    min_date = min(of _diag(*));

    *find first location of smallest value;
    index_min_date = whichn(min_date, of _diag(*));

    *find name of variable at index;
     flag = vname(_diag(index_min_date));
end;
else flag = 'Disease Free';

run;

---

@kelly2016 wrote:

Hello SAS Community,
I'm a relatively new SAS programmer and I really need your help coding this!  I have a dataset that contains 5 columns, each representing a different disease diagnosis date in date format (date9.).  Some columns contain missing values.  I need to compare the diagnosis dates across these 5 columns when the value is not missing and flag or print out the earliest diagnosis date variable. Essentially, to find which disease diagnosis occurred first and flag that. 

Is there an easy way to code this?  See my sample data and expected output below. Thank you so much for your help!!

Data Example:

ID	Diag1	Diag2	Diag3	Diag4	Diag5
1	26-Oct-17		28-Oct-18	26-Mar-19
2	.	.	.	.	.
3	22-Dec-21	.	.	.	.
4	.	.	.	10-May-21	13-Nov-22
5	15-Oct-19	15-Oct-19	.	24-Jun-15	23-Oct-17
6	25-Dec-16	14-Oct-18	24-Nov-22	26-Dec-15	23-Oct-23

I would like the output to be as follows:

ID	Diag1	Diag2	Diag3	Diag4	Diag5	Flag
1	26-Oct-17		28-Oct-18	26-Mar-19		Diag1
2	.	.	.	.	.	Disease Free
3	22-Dec-21	.	.	.	.	Diag1
4	.	.	.	10-May-21	13-Nov-22	Diag4
5	15-Oct-19	15-Oct-19	.	24-Jun-15	23-Oct-17	Diag5
6	25-Dec-16	14-Oct-18	24-Nov-22	26-Dec-15	23-Oct-23	Diag4

---

One way is above, there are others. 

[kelly2016]

Hi Reeza,

Thank you SO much for your super quick response! Your codes worked like magic and it is so concise! I have been struggling with this coding the whole afternoon and I am so grateful for your help! 

[kelly2016]

Hi Reeza,

Thank you SO much for your super quick response! Your codes worked like magic and it is so concise! I have been struggling with this coding the whole afternoon and I am so grateful for your help!

[PaigeMiller]

Are these dates actual numeric SAS dates formatted as 26-Oct-17, or are they character strings?

[kelly2016]

Thank you, PaigeMiler. Yes, these SAS dates are numeric values in my dataset.