[Nick Cox]

#15

1 What it says!

2 Already answered in earlier posts.

[Atif Awad]

As there is a function sqrt() the problem will be something related. It's my guess that one or more of the arguments fed to an expression that includes a sqrt() call is undefined, and Stata is puzzled.

You haven't given us a reproducible example but you're already flagging a dataset that might be problematic. If you

set trace on
set traced 1

and run your code again and show us the few lines of output before it stops, then we will see precisely what the program is choking on.

Hello
I tried tody what you suggested to solve this matter, so far I don't know how to solve this problem?

[Nick Cox]

Atif Awad Sympathy for your problem, but #17 doesn't tell us anything new or specific about your set-up, so it seems difficult to add to previous suggestions.

[Atif Awad]

Hello Cox
As you suggested, I used codes
set trace on set traced 1 below is what i get, do undersand somthing?
unknown function *sqrt()
if "`abs'" != "" {
mat `ABS'[`i',`j'] = abs(`r(rho)')
}
if "`show'" != "" {
mat `COR'[`i',`j'] = `r(rho)'
}
}
else {
qui spearman __e`i' __e`j'
mat `E'[`i',`j'] = (`r(rho)')^2
if "`friedman'" != "" {
mat `F'[`i',`j'] = `r(rho)'
}
if "`show'" !="" | "`abs'"!="" {
qui corr __e`i' __e`j'
if "`abs'" != "" {
mat `ABS'[`i',`j'] = abs(`r(rho)')
}
if "`show'" != "" {
mat `COR'[`i',`j'] = `r(rho)'
}
}
}
}
}

[Nick Cox]

In xtcd (SSC) the only point at which sqrt() is invoked is in this block of code

    mat `E'`name'[`i',`j'] = `r(rho)'*sqrt(`r(N)')
                    mat `Nobs'`name'[`i',`j'] = `r(N)'
                    mat `ABS'`name'[`i',`j'] = abs(`r(rho)')
                    mat `COR'`name'[`i',`j'] = `r(rho)'

If a previous correlation goes well then r(rho) r(N) will be defined but if they are not defined Stata gets puzzled. It seems likely that some other deeper problem is implied but to get Stata to execute the entire command a minimum hack might be

    capture {
           mat `E'`name'[`i',`j'] = `r(rho)'*sqrt(`r(N)')
           mat `Nobs'`name'[`i',`j'] = `r(N)'
           mat `ABS'`name'[`i',`j'] = abs(`r(rho)')
           mat `COR'`name'[`i',`j'] = `r(rho)'
}

except that (1) I haven't tested for this (2) I have no idea what other problems may lurk downstream (3) This is speculation and I don't take responsibility (4) email the author (who has moved since the code was posted on SSC, so you will have to Google for his address

Very similar comments apply to xtcsd (SSC) so far as I can see.

These are community-contributed (user-written) commands. You paid nothing for them and everything hinges on what support the authors may give if you are unable to fix the code yourself.

[Jesse Wursten]

Have you tried using -xtcdf- instead?

[zoni ch]

Hi everyone,

i m facing the same problem of error 133 during the cross sectional dependence test in stata/MP 14.2

xtcsd, pesaran abs

unknown function *sqrt()

r(133);


can anyone help me to resolve the issue please..???

[Nick Cox]

#20 is already the best answer I have.

[deepali kalia]

Hi I am facing a similar issue,. On running xtscd, Pesaran I am getting unknown function *sqrt(). My data is unbalanced panel with N >T.Please suggest

[Jennie Nguyen]

xtcd is from SSC, as you are asked to explain. What you show us has no obvious relation to the problem as the square root function isn't shown.

I think what's happening is that your data don't yield a particular correlation and that's not being trapped by the code.

I'd suggest contacting the author directly (Google for his current address). He's likely to want to see your dataset.

i have the same problem with sqrt()
- mat `E'[`i',`j'] = `r(rho)'*sqrt(`r(N)')
= mat __000003[1,7] = *sqrt(0)
unknown function *sqrt()
if "`abs'" != "" {
mat `ABS'[`i',`j'] = abs(`r(rho)')
}
if "`show'" != "" {
mat `COR'[`i',`j'] = `r(rho)'
}
}
else {
qui spearman __e`i' __e`j'
mat `E'[`i',`j'] = (`r(rho)')^2
if "`friedman'" != "" {
mat `F'[`i',`j'] = `r(rho)'
}
if "`show'" !="" | "`abs'"!="" {
qui corr __e`i' __e`j'
if "`abs'" != "" {
mat `ABS'[`i',`j'] = abs(`r(rho)')
}
if "`show'" != "" {
mat `COR'[`i',`j'] = `r(rho)'
}
}
}
}
}
Can anyone explain and give me solution to fix this

[Nick Cox]

Jennie Nguyen You're already quoting the explanation of your problem. xtcd is a community-contributed command with this known bug or limitation mentioned several times in this thread. The solutions are (1) a user with enough skill can clone the command and fix the code (2) people can contact the author (3) people can seek an alternative command. I don't work in this territory and so cannot advise or help further. Sorry if that is disappointing, but it's one thing to spot a problem and quite a different thing to know what a command should do instead. (I am not an econometrician.)

[Daniel Feenberg]

The 3rd line of the trace in #25 shows that the problem is that "`r(rho)'" is evaluating to an empty string, the problem isn't with "sqrt". With the macro evaluating to nothing, the Stata parser thinks the right hand side of the assignment is "*sqrt(0)" (notice the *) instead of something times sqrt(0). This is probably something for the author of the procedure to deal with, but no harm looking at the .ado to see if you can tell more.

[Brian Munene]

Hi i faced the same problem rescently unknown function *sqrt() while using the stata command xtcd i later solved it as i checked the data i was working with had fewer observations that i had dropped as they were missing retaining them and proceed solved the problem as it works on a larger sample

[Brian Munene]

Hi i faced the same problem recently unknown function *sqrt() while using the stata command xtcd i later solved it as i checked the data i was working with had fewer observations that i had dropped as they were missing retaining them and proceed solved the problem as it works on a larger sample

[Mohammad Hossain Ali]

As there is a function sqrt() the problem will be something related. It's my guess that one or more of the arguments fed to an expression that includes a sqrt() call is undefined, and Stata is puzzled.

You haven't given us a reproducible example but you're already flagging a dataset that might be problematic. If you

set trace on
set traced 1

and run your code again and show us the few lines of output before it stops, then we will see precisely what the program is choking on.

Dear expert! Nick Cox
I am facing a similar problem and getting the following output before it stops. It would be very helpful if you could suggest some ways out.