I have a dataset of fund holdings as attached. It is currently in a wide format. I would like to change it to a long format where there is one date column and one number_of_share column. Attached is  I tried using the following code to transpose but it does not work. It says No variables to tranpose. I tried with a different dataset and it works proc transpose data=sample2 out=sample_long; by fund_ticker constituent_ticker name Detail_Holding_Type ; run; ... View more

I am doing an event study. I have a dataset containing trading days and the main dataset. The following code adds 10 trading days before and 100 days after each day in the main dataset. However, it is very slow. Is there another way to achieve the same task or How can I improve efficiency? data tradingdays; set wrds.tradingdays;run; options intervalds=(tradingdays=tradingdays); proc sort data=main_sample;by stock date;run; data main_sample(rename=(date=date_t0 date2=date)); set main_sample; by stock date; if first.date; format date2 YYMMDD10.; do t=-10 to 100; date2=intnx('tradingdays',date,t) ;output; end; run; Edit: Added a sample of the datasets. An example of my main data looks like this: stock date AAA 20190123 BBB 20181110 CCC 20100701 What I would like to achieve is something like this: stock date_t0 date t AAA 20190123 20190117 -3 AAA 20190123 20190118 -2 AAA 20190123 20190122 -1 AAA 20190123 20190123 0 AAA 20190123 20190124 1 AAA 20190123 20190125 2 AAA 20190123 20190128 3 BBB 20181109 20181106 -3 BBB 20181109 20181107 -2 BBB 20181109 20181108 -1 BBB 20181109 20181109 0 BBB 20181109 20181112 1 BBB 20181109 20181113 2 BBB 20181109 20181114 3 CCC 20100701 20100628 -3 CCC 20100701 20100629 -2 CCC 20100701 20100630 -1 CCC 20100701 20100701 0 CCC 20100701 20100702 1 CCC 20100701 20100705 2 CCC 20100701 20100706 3 ... View more