Dealing with missing values in PROC EXPAND

somebody — Thu, 21 Nov 2019 11:56:04 GMT

I have a dataset of daily observations. I am using thhe PROC EXPAND to compute rolling standard deviation over the past 252 days for each stock. For some stocks, there are missing values. How can I tell SAS not to use these missing values? I have an option to set it to 0 (using TRANSFORMIN=(SETMISS 0)) but what if I don't want that? For example, if there are 100 missing values over the past 252 days, I would like to use only 152 obs with values.

PROC EXPAND DATA=daily_returns OUT=daily_returns2;

CONVERT return=std /TRANSFORMIN=(SETMISS 0) TRANSFORMOUT=(MOVSTD 252);

by stock;

RUN;

Re: Dealing with missing values in PROC EXPAND

PeterClemmensen — Thu, 21 Nov 2019 12:01:00 GMT

That is the default behavior of PROC EXPAND. Read the Missing Values section of the PROC EXPAND: Transformation Operations Documentation.

topic Re: Dealing with missing values in PROC EXPAND in SAS Programming

Dealing with missing values in PROC EXPAND

Re: Dealing with missing values in PROC EXPAND