Closed ppaspp closed 6 years ago
Can you turn on trace
in pargegen
and show the log so we can see what exactly is causing the syntax error.
Log file after including set trace on
in pargegen
.
Huh, I don't know what the reldif
Stata function is (it's not in my setup). It's possible the syntax error is actually inside reldif
so you could try increasing the trace depth until it doesn't go any deeper on the error. Do it work in serial (non-parallel) fashion?
hm, I think trace depth is set to infinity according to the stata manual. At least changing it to 10
doesn't give me more information.
Regarding your second question: Do you mean I should run the program with parallel setclusters 1, f
? Then the same error appears. Runing gegen
outside the pargegen program also works.
My personal impression is that the local level
, which is defined by regress, is somehow altered by parallel in combination with gegen. According to the stata manual level(95)
is the default option but in line 4302 of the log file 95,0
shows up. Even when I define the regress option , level(95)
explicitly, the same error is shown.
I was wondering if you could run the program without parallel
at all. Something like bootstrap, reps(50): pargegen price weight foreign rep78
?
Got it. Running pargegen price weight foreign rep78
without any bootstrapping returns the same error, so I guess I should report the error to github/stata-gtools instead
Good luck
Preliminaries
Before submitting an issue, please check (with
x
in brackets) that you:Expected behavior and actual behavior
Use
gegen
from gtools ado instead ofegen
with parallel bs Parallel child processes stop with error. Log:Steps to reproduce the problem
Minimal nonsense example (runs fine when
gegen
is replaced withegen
):System information
Some relevant information
Output from
creturn list
: