Proc expand returns values for empty observations

somebody — Mon, 14 Oct 2019 01:44:23 GMT

I am using PROC EXPAND to compute daily returns. Theoretically, If the variable is empty, the return should be empty (or at least 0). However, PROC EXPAND gives value for the return of empty observations. Why is it? and How to fix it?

Here is my code:

	proc expand data=data_in out=data_out(drop=time);
		by fund_ticker constituent_ticker;
		convert total_shares_held=etf_growth/ transformout=(ratio 1 -1);
	run;

And here is the result:

Re: Proc expand returns values for empty observations

mkeintz — Mon, 14 Oct 2019 02:56:17 GMT

I believe PROC EXPAND defaults to interpolation using METHOD=SPLINE, which I believe is projecting the spline backwards accross the leading observations (and other observations) with missing values.

Try METHOD=NONE in either the PROC EXPAND statement or the CONVERT statement.

topic Re: Proc expand returns values for empty observations in SAS Programming

Proc expand returns values for empty observations

Re: Proc expand returns values for empty observations