I touched on this earlier but think I got off on the wrong path. I have 125K plus records that vary in the place of service. An example is: id                plc of service             dos                         disch 111             hospital                     11dec2010              15dec2010 111             hospice                    15dec2010              17dec2010 111             rehab                       17dec2010              30dec2010 111            hospital                     12nov2011              15nov2011 222            hospital                     01jan2010               10jan2010 333            hospital                     01feb2011               07feb2011 444            hospital                     01jun2010               07jun2010 444            rehab                         07jun2010              15jun2010 444            nursing                     15jun2010               17jun2010 444            hospital                     17aug2011              30aug2011 I need to say that if anything other than hospital shows up, then look at the disch and dos. If the dos is the same as the hospital disch date then all those records go into one file where. So my outcome I want to be: Hospital table: id                plc of service             dos                         disch 111            hospital                     12nov2011              15nov2011 222            hospital                     01jan2010               10jan2010 333            hospital                     01feb2011               07feb2011 444            hospital                     17aug2011              30aug2011 Other table: id                plc of service             dos                         disch 111             hospital                     11dec2010              15dec2010 111             hospice                    15dec2010              17dec2010 111             rehab                       17dec2010              30dec2010 444            hospital                     01jun2010               07jun2010 444            rehab                         07jun2010              15jun2010 444            nursing                     15jun2010               17jun2010 I started my code and know I need a subquery but not sure I am doing the subquery right. PROC SQL; CREATE TABLE ReAdm.Claims_inpt2 AS (SELECT DISTINCT A.* (SELECT * B.ADMIT_DT FROM ReAdm.Claims_inpt WHERE A.CLM_PL_OF_SRVC_DESC <> 'INPATIENT HOSPITAL' AND B.ADMIT_DT = A.ADMIT_DT AND B.MBR_SYS_ID = A.MBR_SYS_ID) AS OTHERADM) FROM ReAdm.Claims_inpt WHERE A.CLM_PL_OF_SRVC_DESC = 'INPATIENT HOSPITAL'); RUN; The above should give me something like: id                plc of service             dos                         disch                           otheradm 111             hospital                     11dec2010              15dec2010                   15dec2010 111             hospice                    15dec2010              17dec2010                    111            hospital                     12nov2011              15nov2011 222            hospital                     01jan2010               10jan2010 333            hospital                     01feb2011               07feb2011 444            hospital                     01jun2010               07jun2010                     07jun2010 444            nursing                       07jun2010              15jun2010 444            hospital                     17aug2011              30aug2011 Then from there I can take and do another query to isolate only the hospital and only the otheradm. ... View more

I know how to do a simple update query to update if I have the data and am just entering. But not sure how to construct if I am updating fields from another table. PROC SQL; UPDATE ReAdm.TINCHI SET MKT_NM = (Readm.Market_mkt_nm) where (readm.tinchi_mkt_nm) is not null; RUN; ... View more

I am trying to format a max query for the date. My code is: proc sql; create table lib.market as (select distinct id, max(eff_dt) as eff_dt from lib.cleaning group by id); run; The table has in it: id           eff_dt 111        1/1/1990 111        1/1/1991 111        1/12/1991 111        1/15/2010 222        1/1/1991 222        1/1/1993 222        1/1/2011 It returns the following: id            eff_dt 111         -21914 222         13419 How do I make it appear with the dates. I am sure it is some format but not sure. ... View more

I have a proc report code that has a ton of code in it but in my define statement I want to say (F/C) or show what columns I am using to divide by. Here is the example of the first part of my code: proc report nowd data=implant2.finalFI split='/'; columns sys_drg_cd TotalClaims TotalAllowed TotalEPD Total_Claims Total_Allowed Total_EPD perc1 Total_Imp_Claims Total_Imp_Allowed Total_Imp_EPD perc2 TtlClaims TtlAllowed TtlEPD perc3 Ttl_Claims Ttl_Allowed Ttl_EPD perc4 imp_alloc_allw_amt perc5 imp_epd_calc_allw_amt perc6 perc7; define sys_drg_cd/group 'DRG Code' missing ; define totalclaims/group 'Total FI/Claim Frequency/All Claims' center missing; define totalallowed/group 'Total FI/Allowed Amount/All Claims' format=dollar20.2 center missing; define Totalepd/group 'Total FI/EPD Amount/All Claims' format=dollar20.2 center missing; define total_claims/group 'Total FI/Claim Frequency/Implant Claims' center missing; define total_allowed/group 'Total FI/Allowed Amount/Implant Claims' format=dollar20.2 center missing; define Total_epd/group 'Total FI/EPD Amount/Implant Claims' format=dollar20.2 center missing; define perc1/computed format=percent10.2 'FI/Percent of All Claims/(F/C) ' center missing; compute perc1; perc1=_c5_/_c2_; endcomp; If I put in the () and the / it splits the column because I have defined split. I was trying to figure out how to put in the actual division sign. ... View more