-esttab- -foreach- looping: How do I -replace- on the first loop, but -append- on each subsequent loop?

[Amy Ng]

Hi Statalist,

I'd like to write a -foreach- loop that runs a number of regressions (let's say I want to run `z' number of regressions), then creates a new *.rtf file (replacing any previous *.rtf output) which has all `z' regression results in the same file. I'm thinking I need to -replace- on the first iteration, and then -append- on all subsequent iterations.

Here's the code format I'm working with, can anybody help where it says XXXreplace/appendXXX?

counter=1
foreach indep_var in ///
"y1" ///
"y2" ///
"y3" ///
... ///
"yz" ///
    {
        reg `indep_var' x
        estimates store m`counter'
        local counter=`counter'+1
    }
esttab m* using "output.rtf", XXXreplace/appendXXX

Would you use the code below? Do I need to use -esttab- twice? The reason I ask is that my -esttab- is a sprawling multi-line command with many options, and I'd rather not have all of those replicated in the code twice. And for the second -esttab- command, is there an easy way to get Stata to -esttab- all models (m1 m2 m3 m4 ... mz) except m1, when I don't always know what the value of `z' is?

Thank you.

counter=1
foreach indep_var in ///
"y1" ///
"y2" ///
"y3" ///
... ///
"yz" ///
    {
        reg `indep_var' x
        estimates store m`counter'
        local counter=`counter'+1
    }
esttab m1 using "output.rtf", replace
esttab m2-mz using "output.rtf", append

[Nils Enevoldsen]

First remove the existing output.rtf. Then use esttab *, append.

On another note, instead of using a counter, replace estimates store with eststo and it's prefix option.

[Amy Ng]

Thank you, Nils!

So perhaps

capture erase "output.rtf"

at the beginning?

Why didn't I think of that?! Thank you.

[Nils Enevoldsen]

That's correct. I should have spelled that out.

Simple solutions are not always obvious solutions.