Exclude observations with out of range values

[Hdababs]

Greetings,

I need to exclude observations that include out of range dates. The column names I am trying to evaluate are identical but with a different prefix. In the code below, I am attempting to include the desired date range which supposed ignore date values  greater and less than 2022.

Regards

data reform; set adj_both;
Where "01JAN2022"d <= see_date1,see_date2,see_date3,see_date4,see_date5,......see_date310, <= "31DEC2022"d;
array fills(*) see_date1 - see_date&pdm_count.;
array filldayz(*) Dayz_Supply1 - Dayz_Supply&pdm_count.;
do i = 1 to (&pdm_count - 1) while (fills(i) ^= .);
if fills(i+1) < (fills(i) + filldayz(i)) then do;
fills(i+1) = (fills(i) + filldayz(i));
end;
end;
drop i;
run;

[PaigeMiller]

---

@Hdababs wrote:

Greetings,

I need to exclude observations that include out of range dates. The column names I am trying to evaluate are identical but with a different prefix.

---

What does "identical but with a different prefix" mean? Show us examples. Or do you mean "identical but with a different SUFFIX"?

Where "01JAN2022"d <= see_date1,see_date2,see_date3,see_date4,see_date5,......see_date310, <= "31DEC2022"d;

As you probably know already, the above is not correct SAS syntax, even if you remove the dots ......

But its not even clear what you intend. Is your intention that if any ONE of these 310 variables is not in the year 2022, then the record is deleted? Or do they ALL have to be not in the year 2022 for the record to be deleted? Or something else?

Are these see_date variables numeric or character (according to PROC CONTENTS)? What are typical values of these variables?

[Hdababs]

Thank you. Identical column names: see_date1 see_date2 see_date3.... The prefix are the numbers following the column name. I am calling them prefix. My intention that if any ONE of these 310 variables is not in the year 2022, then the record is deleted. This is a sample of the output. Rows with red font must be deleted.

see_date1	see_date2	see_date3	see_date4	see_date5	see_date6	see_date7	see_date8	see_date9	see_date10
22-Feb-22	23-May-22	21-Aug-22	19-Nov-22	.	.	.	.	.	.
15-Jan-22	15-Feb-22	17-Mar-22	16-Apr-22	1-Sep-22	30-Nov-22	28-Feb-23	.	.	.
18-Jan-22	18-Apr-22	.	.	.	.	.	.	.	.
18-Jan-22	18-Apr-22	.	.	.	.	.	.	.	.
18-Feb-22	19-May-22	.	.	.	.	.	.	.	.
3-Feb-22	4-May-22	.	.	.	.	.	.	.	.
27-Feb-22	29-Mar-22	28-Apr-22	28-May-22	27-Jun-22	.	.	.	.	.
29-Mar-22	17-May-22	15-Aug-22	15-Nov-22	13-Feb-23	.	.	.	.	.
9-Feb-22	14-Mar-22	14-Apr-22	14-Jul-22	16-Oct-22	14-Jan-23	.	.	.	.
18-Mar-22	21-Jun-22	19-Sep-22	19-Dec-22	19-Mar-23	.	.	.	.	.
5-Apr-22	4-Jul-22	2-Oct-22	31-Dec-22	31-Mar-23	.	.	.	.	.
31-Mar-22	1-Jul-22	29-Sep-22	28-Dec-22	28-Mar-23	.	.	.	.	.

[PaigeMiller]

---

@Hdababs wrote:

Thank you. Identical column names: see_date1 see_date2 see_date3.... The prefix are the numbers following the column name. I am calling them prefix. My intention that if any ONE of these 310 variables is not in the year 2022, then the record is deleted. This is a sample of the output. Rows with red font must be deleted.

---

Ok, thank you, this is a suffix, not a prefix.

So let's think about this ... if any ONE of these 310 variables is not in the year 2022 then we want to delete the record. Instead of using some brute force programming, we can summarize the dates in these variables by determining the maximum and minimum date. If the maximum or minimum is not in 2022, we would delete the record.

So

data want;
    set have;
    if min(of see_date:)<'01JAN2022'd or ... 
         then delete;
run;

I have left part of the code as ... because it is your homework assignment to fill in that part.

Also, the above code will only work if the variables are numeric, and represent actual SAS date values (which are the integer number of days since 01JAN1960). If the variables contain any other representation of the dates, this will not work. I did ask in my earlier post if these values were numeric or character, I don't see an answer.

[Hdababs]

Thank you. I will report back how it goes.

[ballardw]

I might use, just to make it clear that the intended logic was based on Year, so when I see the code 9 months later I have a chance of remembering that was the stated limit.  The specific date is more flexible in general though.

data want;
    set have;
    if year(min(of see_date:) )< 2022 or ... 
         then delete;
run;

[Hdababs]

Greetings,

I was able to figure out the issue. There was a macro variable retrieving previous year value. The reporting period is 2022 and the previous year is 2021. Since I am running the code in 2023, I have to indicate that the previous reporting year is 2021.