Poor error message when a function is supplied as inits

[mhollanders]

Hi there,

I'm not sure if this issue belongs here but I'll give it a go. I've written some models and they run fine in JAGS using jagsUI from R. However, fitting the same model in Multi/OpenBUGS gives me the error:

Error in datalist[[i]]: subscript out of bounds

Google hasn't told me how to solve this... I'd like to use MultiBUGS to speed up running the models but for some reason this is happening.

Cheers,
Matt

[ghahnghahn]

Do you have a code snippet? I could try taking a look.

[mhollanders]

Hi there, I just recently cleaned up that code and compiled some data files for someone else to look at. Would you like me to forward you the email + files I sent to him? Cheers, Matt

…

On Wed, Aug 26, 2020 at 9:54 AM ghahnghahn ***@***.***> wrote: Do you have a code snippet? I could try taking a look. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <https://github.com/MultiBUGS/MultiBUGS/issues/54#issuecomment-680325229>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQAE6RL2CYKLGUMA6I6WLUTSCRFMPANCNFSM4PS4GRAA> .

[ghahnghahn]

I don't happen to use JAGS but I have been a BUGS user for some time. If you have something cleaned up that you wanted to post, I can take a look at it. Disclaimer: I am not on the development team, I am just an "average user". (Side note: there is a WinBUGS e-mailing list and it would be common there for people to send out the non-working code snippets for the user group to assist with. The WinBUGS language should be almost identical to the MultiBUGS language from what I understand.)

[mhollanders]

Hmmm... well could you send me your email address so I can send you the code to try out on your end with (Multi)BUGS? :)

…

On Wed, Aug 26, 2020 at 10:29 AM ghahnghahn ***@***.***> wrote: I don't happen to use JAGS but I have been a BUGS user for some time. If you have something cleaned up that you wanted to post, I can take a look at it. Disclaimer: I am not on the development team, I am just an "average user". (Side note: there is a WinBUGS e-mailing list and it would be common there for people to send out the non-working code snippets for the user group to assist with. The WinBUGS language should be almost identical to the MultiBUGS language from what I understand.) — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <https://github.com/MultiBUGS/MultiBUGS/issues/54#issuecomment-680356214>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQAE6RIKG42BUCTXJU5AXHLSCRJODANCNFSM4PS4GRAA> .

[rjbgoudie]

Hi, sorry for the slow reply.

Possibly part of your data is not the shape/format that R2MultiBUGS or R2OpenBUGS is expecting? I would guess that the error you report is coming from formatdata() in R2MultiBUGS or formatdata() in R2OpenBUGS (which are the same function).

If your data/model is public, feel free to attach to this issue; or email me (follow the link in my profile).

[rjbgoudie]

The error message is due to using the following as inits<-function(){list()}. Either R2MultiBUGS should accept a function as inits to match jagsUI, or should at least give a better error message when a fucntion is used