02-06-2012 01:03 PM
Remove some headers or all headers?
For multiple columns you have to use the _c#_ notation, where # is the number of the column in the report.
02-06-2012 01:19 PM
proc report data=xyz out=pqrs;
col SOLID Group Zone nca,(segment,(tot_cnt FD_cnt fd_pen));
define SOLID /r group;
define Group / "Mega zone" Group;
define Zone / Group;
define nca / across;
define segment / across;
define tot_cnt / analysis;
define fd_cnt / analysis;
define fd_pen / computed format=percent6.0;
rbreak after / ol summarize;
fd_pen = _c5_/_c4_;
in the bove report iam not getting the desired ouput for the computed column???????????
As asked earlier how can we remove the headers created becuase of usage of multiple across functions????????????
02-06-2012 01:58 PM
For TOT_CNT, FD_CNT and FD_PEN, if you want to remove their headers, you would use ' ' (quote space quote) on the DEFINE statement for each variable.
For the ACROSS items, NCA and SEGMENT, ' ' (quote space quote), again, would blank out the headers for those items in the header area.
As for your question about the COMPUTED column, since you are computing multiple occurences of FD_PEN, since that will be computed UNDER each combination of your ACROSS item, then you would have something like (not tested with your data -- so the numbers may need to change):
_c6_ = _c5_/_c4_;
_c9_ = _c8_/_c7_;
For each occurenct of FD_PEN you need to calculate under the combinations of NCA and SEGMENT. To find out, for sure the absolute column names that REPORT is creating for each of your ACROSS variables, look at the OUT= file with PROC PRINT and you should be able to figure out the right absolute column numbers to use.