How to drop last 0

[Jenna Grabey]

I have a dummy dataset that looks like the following:

* Example generated by -dataex-. For more info, type help dataex
clear
input str7 varlist
"A111110"
"A11110"
"A1110"  
"A110"   
"A10"    
"A10110"
"A1010"  
"A111000"
"A1100"  
end

And I am trying to use the below code in order to drop the last 0 from each observation. I took this line of code from here: Stata | FAQ: Removing leading or trailing zeros from string variables

replace varlist = usubstr(varlist, 1, ustrlen(varlist) − 1) if usubstr(varlist,−1,1) == "0"

However when I run this I get the following error message:


unknown function ()
r(133);

Any advice would be much appreciated.

BW,
Jenna

[Nick Cox]

It works if you retype the expression. I think you got some exotic character that looks like a space but isn't a plain space tangled up in your code.

[Jenna Grabey]

Great, thank you Nick. That was an easier fix than expected.

[Nick Cox]

Has that code been anywhere near a word processor such as MS Word?

[Jenna Grabey]

No, I copied it straight from the website into Stata.

[Leonardo Guizzetti]

NIck's hunch was in the right direction.

The corrected code is below. The issue is that the minus signs were not the traditional ASCII character 45 [=char(45)], but a Unicode minus sign (\u2212). Stata will not interpret these characters synonymously, and thus gives an error.

replace varlist = usubstr(varlist, 1, ustrlen(varlist) - 1) if usubstr(varlist,-1,1) == "0"

[Jenna Grabey]

Leonardo Guizzetti thank you but I had it sorted already.

[Nick Cox]

So, when I look at the link in #1 it has my name on it!

But what you're citing is a StataCorp edit. I didn't suggest or change the function names to what they are now (supporting Unicode too).

StataCorp are generous about attributing work to people outside the company, but then company changes are not attributed even if your work is completely rewritten. I just wrote some text a while ago, with Jamie Griffin who I think asked a question on Statalist, and StataCorp did all the rest in terms of HTML mark-up and so forth.

In short, any non-spaces were almost certainly introduced or produced somehow by a browser. The original authors did not intend to bite you, as you will appreciate.

EDIT: Broadly consistent with #6 from Leonardo Guizzetti even if the details were quite wrong!

[Leonardo Guizzetti]

Leonardo Guizzetti thank you but I had it sorted already.

Of course. My post was intended to help those future readers who may encounter a similar problem and come to the forum in search of a solution.