R help - Feb 2013 - Count function calls

Dear R-users,

I have the following problem: I am running the function 'pvcm' from the
'plm' Panel Data package. Inside this function 'solve' is called
and gives for a certain individual data series an exception because of
singularity. I would like to know which individual data series causes this
error. I tried to debug it, but this is truly painful, as solve is called inside
of 'lapply' and there are over 5,000 individual data series in the
panel.

Now, what I would like to do is to count the calls to 'solve' inside the
function, so I can see, where the function throws the exception. I tried to use
'trace' with a count variable, but I have no clue how to define a global
variable to be used by trace and updated at every call.....is there another
approach?


Best

Simon

This is where <<- is helpful:
   > N <- 0 ; trace(solve, quote(N <<- N + 1), print=FALSE)
   Tracing function "solve" in package "base"
   [1] "solve"
   > lapply(3:0, function(i)solve(diag(i,3), 1:3))
   Error in solve.default(diag(i, 3), 1:3) : 
     Lapack routine dgesv: system is exactly singular: U[1,1] = 0
   > N
   [1] 4

You can also set
   options(error=recover)
to look at the state of things when the error occurs.

Bill Dunlap
Spotfire, TIBCO Software
wdunlap tibco.com

> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at
r-project.org] On Behalf
> Of Simon Zehnder
> Sent: Tuesday, February 26, 2013 2:53 AM
> To: r-help at r-project.org help
> Subject: [R] Count function calls
> 
> Dear R-users,
> 
> I have the following problem: I am running the function 'pvcm' from
the 'plm' Panel Data
> package. Inside this function 'solve' is called and gives for a
certain individual data series
> an exception because of singularity. I would like to know which individual
data series
> causes this error. I tried to debug it, but this is truly painful, as solve
is called inside of
> 'lapply' and there are over 5,000 individual data series in the
panel.
> 
> Now, what I would like to do is to count the calls to 'solve'
inside the function, so I can
> see, where the function throws the exception. I tried to use
'trace' with a count variable,
> but I have no clue how to define a global variable to be used by trace and
updated at
> every call.....is there another approach?
> 
> 
> Best
> 
> Simon
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

Hello. 
Another thing you may want to do depends on whether you are using
model="within" (the default) or model="random". 

In the first case, pvcm() estimates separate regressions, so you just
need to loop lm() on individual indices to spot where it fails.

In the second case, what you may want to do is try a similar estimator:
pmg(..., model="mg"), which is an unweighted version of Swamy's
estimator in pvcm() and a simpler function to read and modify, possibly
using the global assignment operator '<<-' as already suggested by
William to output diagnostics.

Best,
Giovanni
> -----Original Message-----
> From: r-help-bounces at r-project.org
[mailto:r-help-bounces at r-project.org] On Behalf
> Of Simon Zehnder
> Sent: Tuesday, February 26, 2013 2:53 AM
> To: r-help at r-project.org help
> Subject: [R] Count function calls
> 
> Dear R-users,
> 
> I have the following problem: I am running the function 'pvcm' from
the 'plm' Panel Data
> package. Inside this function 'solve' is called and gives for a
certain individual data series
> an exception because of singularity. I would like to know which
individual data series
> causes this error. I tried to debug it, but this is truly painful, as
solve is called inside of
> 'lapply' and there are over 5,000 individual data series in the
panel.
> 
> Now, what I would like to do is to count the calls to 'solve'
inside
the function, so I can
> see, where the function throws the exception. I tried to use
'trace'
with a count variable,
> but I have no clue how to define a global variable to be used by trace
and updated at
> every call.....is there another approach?
> 
> 
> Best
> 
> Simon
> 
> ______________________________________________
?
Ai sensi del D.Lgs. 196/2003 si precisa che le informazi...{{dropped:12}}