Import selective records

[Babloo]

I've a CSV file which has last one year (Aug'15 to Jul'16) data. Now my requirement is to import only last 3 months (Jul'16 , Jun'16 & May'16) records without hardcoding. File has a date variable which has values like 21JUN2016 , 20JUL2016 , 31DEC2015 alongside other variables.

Appreciate if someone guide me with the code.

[kumarK]

can you post some sample have dataset and what columns you want that would be helpful to give better solution. Thanks

[Babloo]

Since it is a interview question I don't have any example dataset. I need
to import all the variables including the key variable date.

##- Please type your reply above this line. Simple formatting, no
attachments. -##

[Reeza]

@Babloo Shouldnt you complete interview questions on your own? It's not a fair representation of your skill set otherwise. 

[jklaverstijn]

How about this:

proc sql;
    create table want as
        select * from have
        where month(datevar) + 3 > (
           select max(month(datevar) from have));
quit;

Hope this helps,

- Jan.

[Ksharp]

You'd better post some data to let us test code:


proc sql;
create table want as
 select *
  from sashelp.air
   having date ge intnx('month',max(date),-2) ;
quit;

[Babloo]

Thanks for the code.

Your code subsetting the data from master SAS dataset (sashelp.air). But my question is to creating a master datset with subset of records. i.e. I will need to create a SAS dataset by importing only selective records.

[Ksharp]

?? Once you import that CSV file correctly and there is a DATE type variable date , you can use my code ,just replace SASHELP.CLASS with your HAVE dataset.

[Babloo]

I need to filter the records while importing not after importing

##- Please type your reply above this line. Simple formatting, no
attachments. -##

[ballardw]

---

@Babloo wrote:
I need to filter the records while importing not after importing

---

Good luck with that. If you are supposed to determine the last three months from the contents of the data file without reading it first then I suspect that you may have partially misunderstood the question.

Generically I would have looked at

1) read the data

2) determine the last date in the data set

3) then build a filter and subset the original data.

Or possibly the purpose of the question would be to get you to request clarification of the process.

[Reeza]

---

---

I suspect that you may have partially misunderstood the question.

Interview questions are meant to test both understanding and technical ability so this is a good question according to that. 

And yes, it's being misunderstood.

The solution is a single line of code (WHERE clause) in the data input step based on todays date and intnx function.