Suppress warnings in l1()?

[guido-s]

Hi,

We just added a new method to the development version of R package netmeta which uses l1() inside netcontrib() to calculate network contributions (guido-s/netmeta@2bf0066).

Due to the network structure we have many zeros resulting in loads of warnings like
warning: solve(): system is singular; attempting approx solution
and
warning: solve(): system is singular (rcond: 1.38373e-39); attempting approx solution

Would it be possible to add an argument to l1() to suppress these warnings?
Or is there another possibility to get rid of the warnings?

Best,
Guido

P.S. See, e.g., the output of the following commands:
library("netmeta")
example(netcontrib)
netcontrib(net1, method = "cccp")

[bpfaff]

Hi Guido, yes, you can use ctrl(trace=FALSE) as argument optctrl in the call of l1. Best wishes, Bernhard *Von: *Guido Schwarzer ***@***.***> *An: *bpfaff/cccp ***@***.***> *Kopie: *Subscribed ***@***.***> *Datum: *31.07.2023 14:06:42 *Betreff: *[bpfaff/cccp] Suppress warnings in l1()? (Issue #5)

…

Hi, We just added a new method to the development version of R package *netmeta* which uses /l1()/ inside /netcontrib()/ to calculate network contributions ***@***.***[guido-s/netmeta@2bf0066]). Due to the network structure we have many zeros resulting in loads of warnings like warning: solve(): system is singular; attempting approx solution and warning: solve(): system is singular (rcond: 1.38373e-39); attempting approx solution Would it be possible to add an argument to /l1()/ to suppress these warnings? Or is there another possibility to get rid of the warnings? Best, Guido P.S. See, e.g., the output of the following commands: library("netmeta") example(netcontrib) netcontrib(net1, method = "cccp") — Reply to this email directly, view it on GitHub[#5], or unsubscribe[https://github.com/notifications/unsubscribe-auth/ACM4F4HMAUCT4OTYPOHJ5ADXS6MSNANCNFSM6AAAAAA26ETYVY]. You are receiving this because you are subscribed to this thread.[Verfolgungsbild][https://github.com/notifications/beacon/ACM4F4ABH2WEFKEGIM3LYZ3XS6MSNA5CNFSM6AAAAAA26ETYV2WGG33NNVSW45C7OR4XAZNFJFZXG5LFVJRW63LNMVXHIX3JMTHG2BGDFU.gif]

[guido-s]

I should have said that we are already using this argument in our l1() call. Here is the printout for a reduced example (which produces fewer warnings):

> res <- l1(A1, optctrl = ctrl(trace=FALSE))

warning: solve(): system is singular (rcond: 1.90474e-18); attempting approx solution

warning: solve(): system is singular (rcond: 1.90474e-18); attempting approx solution
>
> res <- l1(A1, optctrl = ctrl(trace=TRUE))

warning: solve(): system is singular (rcond: 1.90474e-18); attempting approx solution

warning: solve(): system is singular (rcond: 1.90474e-18); attempting approx solution
Optimal solution found.

Does this mean that the argument ctrl(trace=FALSE) doesn't work as expected?

[bpfaff]

Ups, sorry, no it does work as intended, though I erroneously assumed that warnings() are supressed, too. Having said this, does a prior supressWarnings() is a remedy. Best wishes, Bernhard *Von: *Guido Schwarzer ***@***.***> *An: *bpfaff/cccp ***@***.***> *Kopie: *Bernhard Pfaff ***@***.***>; Comment ***@***.***> *Datum: *31.07.2023 18:06:10 *Betreff: *Re: [bpfaff/cccp] Suppress warnings in l1()? (Issue #5)

…

I should have said that we are already using this argument in our /l1()/ call. Here is the printout for a reduced example (which produces fewer warnings): *> res <- l1(A1, optctrl = ctrl(trace=FALSE)) warning: solve(): system is singular (rcond: 1.90474e-18); attempting approx solution warning: solve(): system is singular (rcond: 1.90474e-18); attempting approx solution > > res <- l1(A1, optctrl = ctrl(trace=TRUE)) warning: solve(): system is singular (rcond: 1.90474e-18); attempting approx solution warning: solve(): system is singular (rcond: 1.90474e-18); attempting approx solution Optimal solution found. * Does this mean that the argument ctrl(trace=FALSE) doesn't work as expected? — Reply to this email directly, view it on GitHub[#5 (comment)], or unsubscribe[https://github.com/notifications/unsubscribe-auth/ACM4F4ANYR7PG3JSIDVQIQDXS7JWZANCNFSM6AAAAAA26ETYVY]. You are receiving this because you commented.[Verfolgungsbild][https://github.com/notifications/beacon/ACM4F4E4KPZYEEM7AAA4PHLXS7JWZA5CNFSM6AAAAAA26ETYV2WGG33NNVSW45C7OR4XAZNMJFZXG5LFINXW23LFNZ2KUY3PNVWWK3TUL5UWJTTC3WKXC.gif]

[guido-s]

Neither suppressWarnings() nor options(warn = -1) suppresses the warnings.

Best wishes,
Guido

[guido-s]

I fixed the problem in guido-s/netmeta@b376499 using

invisible(capture.output(..., type = "message"))

Thanks to @wviechtb to point me at https://stackoverflow.com/questions/49694552/suppress-messages-from-underlying-c-function-in-r .