EViews.com

Posted: **Sat Jan 16, 2016 4:34 am**

Dear All,

Is it possible to order the series.freq (one way tabulation) by count rather than alphabetically by alpha name?

Thanks!

Charlie

Posted: **Sat Jan 16, 2016 1:52 pm**

No :(

Posted: **Fri Feb 12, 2016 11:58 am**

now it is possible:

Code: Select all

subroutine frequencytab(alpha inputalpha)
	freeze(inputalpha_tab) inputalpha.freq(nocum,nolimit)
	!numberunique=@rows(@uniquevals(inputalpha))
	wfcreate(wf=frequencytabulated) u !numberunique
	copy alldatasets::inputalpha_tab frequencytabulated::inputalpha_tab
	alpha uniquevalues = ""
	genr uniquevaluefreq = NA
	for !y = 1 to !numberunique
		uniquevalues(!y) = inputalpha_tab(7+!y,1)
		uniquevaluefreq(!y) =@val(inputalpha_tab(7+!y,2))
	next
	group tabber uniquevaluefreq uniquevalues
	tabber.sort(-1)
	freeze(freqtab) tabber
	copy frequencytabulated::freqtab alldatasets::freqtab
	freqtab.deletecol(a) 1
	wfclose frequencytabulated
	d inputalpha_tab
endsub

Posted: **Tue Feb 23, 2016 3:01 pm**

Better to endogenously pull out the wf name - cleaner:

Code: Select all

subroutine frequencytab(alpha inputalpha)
	%wfnameholder = @wfname
	freeze(inputalpha_tab) inputalpha.freq(nocum,nolimit)
	!numberunique=@rows(@uniquevals(inputalpha))
	wfcreate(wf=frequencytabulated) u !numberunique
	copy {%wfnameholder}::inputalpha_tab frequencytabulated::inputalpha_tab
	alpha uniquevalues = ""
	genr uniquevaluefreq = NA
	for !y = 1 to !numberunique
		uniquevalues(!y) = inputalpha_tab(7+!y,1)
		uniquevaluefreq(!y) =@val(inputalpha_tab(7+!y,2))
	next
	group tabber uniquevaluefreq uniquevalues
	tabber.sort(-1)
	freeze(freqtab) tabber
	copy frequencytabulated::freqtab {%wfnameholder}::freqtab
	wfclose frequencytabulated
	d inputalpha_tab
endsub

EViews.com

order .freq by count

order .freq by count

Re: order .freq by count

Re: order .freq by count

Re: order .freq by count