How to skip the r(2000) error and continue with code in Stata

[Tiago Pereira]

Dear all,

My search on this topic has not been satisfactory:


https://www.statalist.org/forums/for...stfile-command
https://www.stata.com/statalist/arch.../msg01072.html
https://www.stata.com/statalist/arch.../msg00382.html


So, my question is: How can I successfully run the code below and skip the instances in which Stata returns r(2000)? E.g., all observations failed (event=1 for all items) or did not fail (event=0 for all items). In my do file, that simple simulation stops at step 2.

*! step 1
clear
set obs 100
gene trt = runiform()<0.5
gene event = runiform()<0.2
logit event trt

*! step 2
cap drop trt event
set obs 100
gene trt = runiform()<0.5
gene event = runiform()<0.001
logit event trt

*! step 3
cap drop trt event
set obs 100
gene trt = runiform()<0.5
gene event = runiform()<0.999
logit event trt

*! step 4
cap drop trt event
set obs 100
gene trt = runiform()<0.5
gene event = runiform()<0.2
logit event trt

All the best,

Tiago

[Andrew Musau]

help capture

Same way as you have

cap drop

to suppress the error message. You can first check whether the logit runs and then rerun it if it is successful.

cap logit event trt
if _rc==0 {
      logit event trt
}

[David Fisher]

The following should also work, and has the advantages that logit is only run once each time, and that the output will show whether or not the r(2000) error was encountered:

capture noisily logit event trt

[Tiago Pereira]

Dear Andrew and David,

As always, thank you so much for your help and tips. Always extremely helpful.

All the best,

Tiago