numeric and string variables: replace a single value of a numeric variable in a specific index

[Chwen Chwen Chen]

Dear Stata users,

I would like to ask your help. I have a numeric variable named "nmf" and a string variable named "nmf_str".

I want to replace the value of the numeric variable at index 5 with the value of the string variable at the same index.
The numeric variable contains names of funds that I previously converted to a numeric format.
The string variable only contains a value at index 5.

My code is the following:

. local new_value = real(nmf_str[5])

. replace nmf = `new_value' in 5
(1 real change made, 1 to missing)


In the data editor the previous value of the numeric variable at index 5 is now a missing value.

I very much appreciate any help!

Best,

Chwen Chwen

[Nick Cox]

Hard to know what is happening but the use of a local macro is not needed here.

Think of these steps.

I have a pen. I put it in a box. I take it out of the box. Now I have a pen again.

to which a natural question is Why did you fool around with the box there?

In short,

replace nmf = real(nmf_str) in 5

is the way to do that directly.

[Chwen Chwen Chen]

Thank you Nick!

I think that I got lost and stopped using my brain for a while...

I got the message below after running the code line:


. replace nmf=real(nmf_str) in 5
(1 real change made, 1 to missing)

which is the same message I got previously.

I report what the data editor shows in the table below (I had to change the names for confidentiality issues):

	nmf_str	nmf
4		m1
5	s1	.

As one can see, the variable nmf displays a missing value instead of "s1".


I also checked both variables but cannot find anything unusual:


. descr nmf_str nmf

Variable Storage Display Value
name type format label Variable label
--------------------------------------------------------------------------
nmf_str str20 %20s
nmf long %101.0g nmf fund name

.

Why does this happen?
Any suggestion about how I can replace the value at index 5?

Thank you.

Chwen Chwen

[Nick Cox]

Indeed. Stata can't see that the string sm1 really is a number wrapped in string form.

di real("sm1")

returns missing. What did you expect to happen there and why?

[Nick Cox]

In #4 the illustration should be in terms of s1 not sm1 but the principle is the same.

[Chwen Chwen Chen]

I have noticed your answer right now, thank you Nick.