You could get that test by rescaling v3:   data test_in_2; set test_in; v3_myVar = v3 / myVar; run; proc reg noprint data=reg_in_2 outest=reg_out ; model dep = v1 v2 v3_myVar v4; by group_id myvar; t1: test v3_myvar + v4 =0; ods output ParameterEstimates=PE TestANOVA=TA; run; Datasets PE and TA will contain the estimates and test results and also the value of myvar so that you can recalculate the v3 statistics.   (untested) ... View more

@jl2978 wrote: Thanks for pointing out the post! However, the situation is for my data, one "date" has multiple observations and I'll need to use all these observations to calculate covariance. So it'll be difficult for me to specify the rolling window in proc expand, which works on datasets that a date only has one observation.  I'm not clear on how you would use multiple observations on a single date in a rolling window covariance, I don't think you have explained the details. But whatever the method, you would have to program that, and since the link I gave also pointed to some DATA step code, its not unreasonable for you to modify the data step code to handle this. ... View more

data have; input str $40.; want=prxchange('s/-\w\w\s*$//',1,str); cards; TUXEDO PARTNERS INC-AT KEYES HOWARD AUTO DLRHSPS-PX SPRING HILL FORD INC-CH MEGASTAR INC BRANDON CHRY-PLYM INC-TM ; ... View more

The character after PC is NOT a space, but another TAB.  So your attempt to replace the tabs with pipes is wrong since you didn't replace that one.   So the error with your file is a different error than what I said before.  The error is that when a value contains the delimiter then the value must be enclosed in quotes in the delimited file.  It does not matter whether the delimiter in the value already is quoted in the value itself like in your example (although any existing quote characters will need to be doubled up).   So the value APPLE IPAD COMPUTER SYS "AIR/TABLET PC|" should appear as  "APPLE IPAD COMPUTER SYS ""AIR/TABLET PC|""" or have the tab changed to a space APPLE IPAD COMPUTER SYS "AIR/TABLET PC " to work as a valid value in a delimited file.   Most likely the only reason that using CSV instead of tab delimited worked for you is that none of your actual data contains commas while at least this one value does contain a tab.   What is the source of this file? Is it being written by some software? Commercial/public software like a database system? Or custom code?  In either case request that they produce valid delimited files.  If the source system does not produce valid delimited files then they cannot be interpreted by anyone.  It you are using a modern software it should be able to know when to add the quotes around the values.  If not you might be able to ask it to add quotes around everything.  Although that will confusing PROC IMPORT into thinking that everything is a character variable.   You could try counting the quotes on the line and replacing any tabs inside the quotes with spaces, but I am not sure that you can be positive that every embedded tab in your data will actually appear inside of quotes in the value. ... View more