My partial data looks like this: Market goes from 1-50, Brand goes from 1-200, QNUM= BI1ATT01...BI1ATT357, BI2ATT01...BI2ATT200, BI3ATT01...BI3ATT260 AttValue= 0 or 1 SURVEYID QNO CQNO MARKET BRAND LOADMONTH MONREP QTRREP WEIGHT XMKT_WEIGHT CREATEDDATE Qnum AttValue TIME_PERIOD FLAG BAC11 6840348 00006840348 130100006840348 10 4 01Jan2013 1301 1301 1.1147700000000000 12.1398900000000000 04Apr2013 19:53:51 BI1ATT01 0 1/1/13 0 6 6840699 00006840699 130100006840699 10 4 01Jan2013 1301 1301 1.1207900000000000 12.2054200000000000 04Apr2013 19:53:55 BI1ATT01 0 1/1/13 0 10 6840763 00006840763 130100006840763 10 4 01Jan2013 1301 1301 1.1207900000000000 12.2054200000000000 04Apr2013 19:53:56 BI1ATT01 0 1/1/13 0 10 6840677 00006840677 130100006840677 10 4 01Jan2013 1301 1301 1.1207900000000000 12.2054200000000000 04Apr2013 19:53:54 BI1ATT01 0 1/1/13 0 9 6840684 00006840684 130100006840684 10 4 01Jan2013 1301 1301 0.9990600000000000 10.8797300000000000 04Apr2013 19:53:55 BI1ATT01 0 1/1/13 0 8 6840565 00006840565 130100006840565 10 4 01Jan2013 1301 1301 1.1998900000000000 13.0667700000000000 04Apr2013 19:53:53 BI1ATT01 0 1/1/13 0 9 6840801 00006840801 130100006840801 10 4 01Jan2013 1301 1301 1.1207900000000000 12.2054200000000000 04Apr2013 19:54:03 BI1ATT01 0 1/1/13 0 10 6840650 00006840650 130100006840650 10 4 01Jan2013 1301 1301 1.1207900000000000 12.2054200000000000 04Apr2013 19:53:54 BI1ATT01 0 1/1/13 0 10 This code gives me correct correlations: PROC CORR DATA=RESP_ATT_2 OUT=RESP_ATT_2_unwtd NOPRINT ; BY TIME_PERIOD MARKET QNUM NOTSORTED; WHERE MARKET=10 AND QNUM = 'BI1ATT01'; VAR BAC11; WITH attvalue ;RUN; Which is: TIME_PERIOD MARKET Qnum _TYPE_ _NAME_ BAC11 1/1/13 10 BI1ATT01 MEAN 6.976191972 1/1/13 10 BI1ATT01 STD 2.080971423 1/1/13 10 BI1ATT01 N 15919 1/1/13 10 BI1ATT01 CORR AttValue 0.219948207 But if I add another qnum or do it for all values of qnum, it splits (uncorrectly): TIME_PERIOD MARKET Qnum _TYPE_ _NAME_ BAC11 1/1/13 10 BI1ATT01 MEAN 7.046176 1/1/13 10 BI1ATT01 STD 2.04887 1/1/13 10 BI1ATT01 N 4916 1/1/13 10 BI1ATT01 CORR AttValue 0.200917 1/1/13 10 BI1ATT02 MEAN 7.046176 1/1/13 10 BI1ATT02 STD 2.04887 1/1/13 10 BI1ATT02 N 4916 1/1/13 10 BI1ATT02 CORR AttValue 0.357214 2/1/13 10 BI1ATT01 MEAN 6.893767 2/1/13 10 BI1ATT01 STD 2.038903 2/1/13 10 BI1ATT01 N 5535 2/1/13 10 BI1ATT01 CORR AttValue 0.227032 2/1/13 10 BI1ATT02 MEAN 6.893767 2/1/13 10 BI1ATT02 STD 2.038903 2/1/13 10 BI1ATT02 N 5535 2/1/13 10 BI1ATT02 CORR AttValue 0.359607 3/1/13 10 BI1ATT01 MEAN 6.996708 3/1/13 10 BI1ATT01 STD 2.148244 3/1/13 10 BI1ATT01 N 5468 3/1/13 10 BI1ATT01 CORR AttValue 0.228997 3/1/13 10 BI1ATT02 MEAN 6.996708 3/1/13 10 BI1ATT02 STD 2.148244 3/1/13 10 BI1ATT02 N 5468 3/1/13 10 BI1ATT02 CORR AttValue 0.329982 4916 + 5535 + 5468 = 15919 How can I get correlations for all bac11 with attvalue for all qnums in 1 proc corr step? I have been able to do it when each qnum was a separate column. Thanks, in advance!. ... View more

Hello, I have SAS EG on my local computer. I have a profile for local and one for the server. I do all the data processing and saving on the SAS server. But I want to include a program which is on local and run it on the server with a statement like this: %include "C:\Documents and Settings\myname\Desktop\Code\For SAS Server\QTR_START_END.sas"; And I get this error:      WARNING: Physical file does not exist, /sas93/config/Lev1/SASApp/C:\Documents and Settings\myname\Desktop\Code\For SAS Server\QTR_START_END.sas ERROR: Cannot open %INCLUDE file C:\Documents and Settings\myname\Desktop\Code\For SAS Server\QTR_START_END.sas I want to use this %include because I use this program in many SAS EG projects but want to change it only once. Thanks, in advance for any help!    ... View more

I know how to export data from SAS using code and from the file menu but I want to know how to give a client the data with the formats. We use a .dat and .sas file to create the data. Now this dataset has to be converted to a .csv file with supporting documentation for formats. I notice that the formats don't export. Thanks! ... View more

This is a part in my macro:   PROC MEANS DATA=&DSN NOPRINT NWAY; CLASS TIME_PERIOD QNUM aggregate BRAND; VAR FLAG; WEIGHT XMKT_WEIGHT; OUTPUT OUT=&DSN._WTDMEANS SUMWGT=BASE_WTD SUM=CNT_WTD MEAN=PER_WTD ; RUN; For example, I have Markets=1,2,3,4,......60 and for example. aggregate=1002 combines markets (5,6, 13,14). So in the above proc means, I want WEIGHT = WEIGHT (my weight variable ) if aggregate=1,2,....60 but I want it to use XMKT_WEIGHT if it's a combination of markets viz aggregate=1002. How can I tell proc means which weight to use based on a condition. If it helps, aggregate=1 is market=1, aggregate=2 is market=2......upto aggregate=60 is market=60...and then aggregate=1000 is market=1,2,3, aggregate=1001 is market=2, 3, 9, 10 etc which I have in a table: AGGREGATE MARKET      1                    1      2                    2      1002               5      1002               6      1002               13  .......... ETC                      Thanks! ... View more