Filling the series in SAS - sas

Question

I have a data which data which looks something like this
/********************************************************************************/
YYMM Sector
1701 Agriculture
1611 Retail
1501 CRE
/*************/
There is another dataset which looks something like this/*************
Customer_ID YYMM
XXXX 1702
XXXX 1701
XXXX 1612
XXXX 1611
XXXX 1610
XXXX 1510
XXXX 1509
/********************************************************/
So basically I want to mere these two datasets on the basis of YYMM but and merge in the sectors. But since the previous data has only few YYMM all I want to do is copy the sectors till a new yymm is encountered from the first dataset.
So the sector from 1701 to 1612 should be agriculture and the sector from 1611 to 1502 is retail and for any month before 1501 it has to be CRE.
Can you please tell me how to do it?

Answer 1

Here is a SQL based solution (similar to the one proposed by pinegulf).
Let us create test datasets:
data T01;
length Sector $20;
infile cards;
input YYMM_to Sector;
cards;
1701 Agriculture
1611 Retail
1501 CRE
;
run;
data T02;
length Customer_id $10;
infile cards;
input Customer_ID YYMM;
cards;
AXXX 1702
BXXX 1701
CXXX 1612
DXXX 1611
EXXX 1610
FXXX 1510
GXXX 1509
;
run;
We can add a "YYMM_from" column to T01:
proc sort data=T01;
by YYMM_to;
run;
data T01;
set T01;
by YYMM_to;
YYMM_from=lag(YYMM_to);
if _N_=1 then YYMM_from=0;
run;
proc print data=T01;
run;
We get:
Obs Sector YYMM_to YYMM_from
------------------------------------------
1 CRE 1501 0
2 Retail 1611 1501
3 Agriculture 1701 1611
Then comes the join:
proc sql;
create table T03 as
select a.*, b.Sector
from T02 a LEFT JOIN T01 b
on YYMM_from<a.YYMM<=YYMM_to;
quit;
proc print data=T03;
quit;
We get:
Obs Customer_id YYMM Sector
-----------------------------------------
1 DXXX 1611 Retail
2 EXXX 1610 Retail
3 FXXX 1510 Retail
4 GXXX 1509 Retail
5 BXXX 1701 Agriculture
6 CXXX 1612 Agriculture
7 AXXX 1702

Answer 2

Here is a solution with proc format. Since your data is in yymm format you can set the limits logical without the data conversion, but I feel more comfortable with actual dates.
data Begin;
input Customer_ID $ YYMM $;
cards;
XXXX 1702
YYYY 1701
ZZZZ 1612
OOOO 1611
AAAA 1610
FFFF 1510
DDDD 1509
; run;
data with_date;
set begin;
date = mdy(substr(yymm,3,2), 1, substr(yymm,1,2) );
run;
proc format; /*Didn't check the bins too much. Adjust as needed.*/
value sector
low - '1jan2015'd ='lows'
'1jan2015'd < - '1nov2016'd = 'CRE'
'1nov2016'd < - '1jan2017'd = 'Retail'
'1jan2017'd < - high = 'Agriculture'
;
run;
data wanted;
set with_date;
format date sector.;
run;
For more on proc format see http://support.sas.com/documentation/cdl/en/proc/61895/HTML/default/viewer.htm#a002473474.htm