Hey guys,
A quick question related to my rolling correlations discussion in a different subforum. I'm trying to operate directly on proc corr output results and substitute rows or cols or specific ranges of values with missing values.
say i have this output
_TYPE_ _NAME_ ADI ALTR AMAT AMD AMKR
MEAN 0.00 0.00 0.00 0.00 -0.01
STD 0.04 0.04 0.03 0.04 0.05
N 126 126 126 126 63
CORR ADI 1.00 0.52 0.46 0.33 0.28
CORR ALTR 0.52 1.00 0.54 0.22 0.34
CORR AMAT 0.46 0.54 1.00 0.38 0.32
CORR AMD 0.33 0.22 0.38 1.00 0.25
CORR AMKR 0.28 0.34 0.32 0.25 1.00
so i want to find all columns/vars with N less than some dynamic nr, say100, and then populat just the corresponding column and row with missing values. So in the example above, the end result should be like this
_TYPE_ _NAME_ ADI ALTR AMAT AMD AMKR
MEAN 0.00 0.00 0.00 0.00 -0.01
STD 0.04 0.04 0.03 0.04 0.05
N 126 126 126 126 63
CORR ADI 1.00 0.52 0.46 0.33 .
CORR ALTR 0.52 1.00 0.54 0.22 .
CORR AMAT 0.46 0.54 1.00 0.38 .
CORR AMD 0.33 0.22 0.38 1.00 .
CORR AMKR . . . . .
been trying to approach this problem from several angles, but right now other than call R within SAS, i see no way. There should be an elegant SAS-wise solution i think. It's somewhat urgent and the code should be dynamic enough to adjust for different n, several variables with N less than some nr and lots of variables.
Thanks!!
Arsenio