Closed giannilmbd closed 1 year ago
Hi Gianni, thanks for reporting! Which version are you running? This almost certainly is an environment error - basically fwildclusterboot does not find the data because it searches in the wrong env. I have updated this a while back and thought / hoped such errors would no longer occur. If you are not running the most up to date version, could you please reinstall fwildclusterboot
and report if the error vanishes? If not, I will start investigating :) As the package is currently not on CRAN (a long story), you can download from r universe (via install.packages('fwildclusterboot', repos ='https://s3alfisc.r-universe.dev')
).
Best, Alex
Here is a related issue - the discussed error was fixed with version 0.9
=)
Thanks Alex. I've actually installed it today from GitHub (devtools::install_github("s3alfisc/fwildclusterboot")). Should I install it from r-universe instead?. Cheers Gianni
On Sat, Nov 19, 2022 at 3:18 PM Alexander Fischer @.***> wrote:
Hi Gianni, thanks for reporting! Which version are you running? This almost certainly is an environment error - basically fwildclusterboot does not find the data because it searches in the wrong env. I have updated this a while back and thought / hoped such errors would no longer occur. If you are not running the most up to date version, could you please reinstall fwildclusterboot and report if the error vanishes? If not, I will start investigating :) As the package is currently not on CRAN (a long story), you can download from r universe (via install.packages('fwildclusterboot', repos ='https://s3alfisc.r-universe.dev')). Best, Alex
— Reply to this email directly, view it on GitHub https://github.com/s3alfisc/fwildclusterboot/issues/79#issuecomment-1320895382, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHLXB3YLT3ZOKJEA5WQ7P53WJDOTHANCNFSM6AAAAAASFJQC6M . You are receiving this because you authored the thread.Message ID: @.***>
Both should lead to the same outcome =) the advantage of r-universe is that the c++ code is already compiled. Did updating the package fix the error?
No luck. I've re-installed from r-universe but got the same error. Indeed it seems it's looking for the data in the main environment, but the data is available only within the function.
On Sat, Nov 19, 2022 at 6:06 PM Alexander Fischer @.***> wrote:
Both should lead to the same outcome =) the advantage of r-universe is that the c++ code is already compiled. Did updating the package fix the error?
— Reply to this email directly, view it on GitHub https://github.com/s3alfisc/fwildclusterboot/issues/79#issuecomment-1320925902, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHLXB36NQ2ZMVTIHBDP5YM3WJECH5ANCNFSM6AAAAAASFJQC6M . You are receiving this because you authored the thread.Message ID: @.***>
Thanks for reporting. Just as a side comment before I dive into this deeper - it looks like you might be looping over "feols()" for multiple specifications? Have you tried to use fixest's multiple estimation functionality? That should be well supported by fwildclusterboot, see the news log for version 0.9 (I should probably add this to the main vignette). Beyond, I hope I'll find time to take a closer look at this tomorrow =)
Great, thanks! I’ll give a try to the multiple specifications.
On Sat, 19 Nov 2022 at 19:39, Alexander Fischer @.***> wrote:
Thanks for reporting. Just as a side comment before I dive into this deeper - it looks like you might be looping over "feols()" for multiple specifications? Have you tried to use fixest's multiple estimation functionality? That should be well supported by fwildclusterboot, see the news log for version 0.9 (I should probably add this to the main vignette). Beyond, I hope I'll find time to take a closer look at this tomorrow =)
— Reply to this email directly, view it on GitHub https://github.com/s3alfisc/fwildclusterboot/issues/79#issuecomment-1320943446, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHLXB322JTTLERJSUH6SAKLWJENHFANCNFSM6AAAAAASFJQC6M . You are receiving this because you authored the thread.Message ID: @.***>
-- Giovanni Lombardo Bank for International Settlements Basel - CH www.giovannilombardo.ch @.***
Hi @giannilmbd, have you managed to work around the problem above?
Hi, No, not yet Cheers Gianni
On Thu, Nov 24, 2022 at 5:39 PM Alexander Fischer @.***> wrote:
Hi @giannilmbd https://github.com/giannilmbd, have you managed to work around the problem above?
— Reply to this email directly, view it on GitHub https://github.com/s3alfisc/fwildclusterboot/issues/79#issuecomment-1326667785, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHLXB3YBMR44OBMWMKXMZLDWJ6K23ANCNFSM6AAAAAASFJQC6M . You are receiving this because you were mentioned.Message ID: @.***>
Hi @giannilmbd , is this still relevant? If not, I'd like to close the issue :) Best, Alex
Thanks Alex, Not relevant right now. I think you can close the issue. Thanks for asking Gianni
On Fri, 30 Dec 2022 at 15:55, Alexander Fischer @.***> wrote:
Hi @giannilmbd https://github.com/giannilmbd , is this still relevant? If not, I'd like to close the issue :) Best, Alex
— Reply to this email directly, view it on GitHub https://github.com/s3alfisc/fwildclusterboot/issues/79#issuecomment-1367958008, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHLXB32OJX2FTV63NRWY6TLWP3ZX5ANCNFSM6AAAAAASFJQC6M . You are receiving this because you were mentioned.Message ID: @.***>
-- Giovanni Lombardo Bank for International Settlements Basel - CH www.giovannilombardo.ch @.***
Hi, Thanks for making your code available. I've run the following
fit.pst3<-fixest::feols(data =data_2,fml= formula(paste(update(form,.~.-factor(cc)),"|cc",sep='')),panel.id =c("cc","year")) summary(fit.pst3) fixest::etable(fit.pst3) boot_feols <- fwildclusterboot::boottest(fit.pst3, clustid = "cc", param = "M3_ratedif", B = 9999)
and got the error ' Argument 'data' could not be evaluated. Problem: object 'data_2' not found.' data_2 is a data.frame object that is accepted with no issues by feols. Note that I'm calling your function from within another function. ThanksGianni