Create a variable from Stata local

Question

I want to know how to create a variable that stores the value of a macro. The most simple example would be something along the lines of storing the value of i' in a forvalues` loop below.

forvalues i = 1/10 {
dis `i'
}

As far as I know, I can generate variables whose name are in the format name'i' but not directly storing the value of i to another variable. What I want to do ultimately is to create a year variable that stores, obviously, the year of the dataset using a loop:

forvalues i = 2000/2020 {
use dataset`i'
gen year`i' = `i'
*gen year`i' = "`i'" ***this did not work either
}

The above code generates a bunch of

year`i' 

with 0 or NULL values.

Answer 1

The syntax looks fine to me. The error report is thus puzzling. Various comments follow:

One guess is that although your quotation marks look fine here, you may have used different characters. The left and right single quotation marks here are in Stata terms uchar(96) and uchar(39).

Stata does not have a concept of NULL. For a numeric variable, generic missing values are indicated by a dot, period or stop .. I can't see why you would get a variable yet also anything but one of 2000/2020 as its value. (Hence a result of 0 is also mysterious.)

A pitfall with local macros is that code containing them must be executed as a whole; otherwise the definition of a local macro is not visible from statements in which they are referenced. But if this was biting it is hard to see why you would get any results at all.

A common motive for doing something like this is as a step to appending datasets, one for each year. But that is made more complicated by having a different variable name in each case. The new variable should have the same name, say year, in each dataset, not different names year2000 and so on.

• Rei: Yes you are right about that appending datasets and yes i intend for the year variable to be the same across datasets, the code above is mainly for illustrative purpose. After double checking, the code works. Perhaps when I was testing with no data in memory, and hence 0 observations, it did not generate anything because the number of observations is 0 rather than a fault in the code itself.
• Nick Cox: Glad you solved your problem, but it seems that you are saying that your code is not what you tried and not even comparable to what you tried. You started off on CV but on any platform minimal reproducible examples are the ideal