data manipulation question

AllSoEasy — Tue, 06 Aug 2013 16:56:22 GMT

Hello all,

From the parameter output of Proc Model ("Nonlinear OLS Parameter Estimates") I am obtaining the Standard Errors of each of my parameter estimates to append to another file of all my parameter data. Basically the data I get from the ods output is formatted as such:

Parameter	StdErr
A1	0.9174
B1	0.2323
B2	0.2418
B3	0.0778
gamma	0.0145

I would like to manipulate this data to a one row data-set as such:

StdErr_A1	StdErr_B1	StdErr_B2	StdErr_B3	StdErr_gamma
0.9174	0.2323	0.2418	0.0778	0.0145

I have been trying to do this by storing the parameter I'm currently on in a datastep to a macro variable that can I use to concatenate a variable name and assign the StdErr to, as such:

data stderr_out2;

set stderr_out;

call symputx("parm", parameter);

stderr_&parm. = StdErr;

run;

This won't give me the desired end-result yet, but I will at least have the correct variable names with their respective values and I can get to this final result with another datastep. However, this is not working at all, I only get one column added that corresponds to the last parameter, i.e. I will get this:


Parameter	StdErr	stderr_gamma
A1	0.917366246	0.917366246
B1	0.232323077	0.232323077
B2	0.241839023	0.241839023
B3	0.077813946	0.077813946
gamma	0.014471111	0.014471111

Rather than a new column for each parameter, only the column from the last macro is placed in the data set. Am I going about this a completely incorrect way? This seems like it should be fairly simple to do and I've gotten somewhat stumped as to what is going on here. Please if you have any suggestions I would appreciate it a great deal.

Thanks!!

-Ryan

Re: data manipulation question

PaigeMiller — Tue, 06 Aug 2013 16:59:03 GMT

I think you want to use PROC TRANSPOSE, with the ID statement, something like (untested)

PROC TRANSPOSE DATA=whatever OUT=transpose;

var stderr;

id parameter;

run;

Re: data manipulation question

AllSoEasy — Tue, 06 Aug 2013 17:02:07 GMT

Well I feel really silly that it was that easy!! I don't know why I didn't think of using proc transpose with the ID statement for this. Thank you for the quick answer!!

-Ryan

topic Re: data manipulation question in SAS Programming

data manipulation question

Re: data manipulation question

Re: data manipulation question