lme4 upgrade produces error message Error in `[[<-.data.frame`(`*tmp*`, i, value = integer(0))

问题

I have constructed a generalized mixed linear model with lmer (lme4) last week which worked fine:

fit<-lmer(dat$presence~log(dat$SIZE_strict)*dat$Troph_level+log(dat$HAB500EXCL_strict+1)+(1|dat$dataset), family=poisson, REML=FALSE)

After updating the lme4 package however, the model no longer worked, giving the error message:

Error in `[[<-.data.frame`(`*tmp*`, i, value = integer(0)) : 
  replacement has 0 rows, data has 174
In addition: Warning messages:
1: In lmer(dat$presence ~ log(dat$SIZE_strict) * dat$Troph_level +  :
  calling lmer with 'family' is deprecated; please use glmer() instead
2: In checkArgs("glmer", REML = FALSE) :
  extra argument(s) ‘REML’ disregarded

Using glmer instead of lmer did not solve the first error message. Any suggestions would be very much appreciated!

Here are some of the data (called dat):

Troph_level presence    dataset SIZE_strict HAB500EXCL_strict
carnivorous 2   1   46155   26005
carnivorous larvae  0   1   46155   26005
phytophagous    2   1   46155   26005
phytophagous    0   3   195295  360882
carnivorous 0   3   195295  360882
carnivorous larvae  0   3   195295  360882
phytophagous    4   2   18272   21169
carnivorous larvae  0   2   18272   21169
carnivorous 1   2   18272   21169
carnivorous 1   2   24964   26745
carnivorous larvae  0   2   24964   26745
phytophagous    4   2   24964   26745
phytophagous    5   2   6220    12543
carnivorous larvae  0   2   6220    12543
carnivorous 1   2   6220    12543
phytophagous    0   3   102633  12198
carnivorous larvae  0   3   102633  12198
carnivorous 0   3   102633  12198
phytophagous    0   3   2092    291439
carnivorous larvae  1   3   2092    291439
carnivorous 0   3   2092    291439
phytophagous    3   5   80410   0

Many thanks in advance! Toos van Noordwijk

回答1:

This is a buglet in the new version of lme4 (to be fixed in an upcoming patch release) that requires grouping variables to be findable within the data frame specified by the data argument. In general things will work more smoothly (and your code will be easier to read) if you use the data argument and eschew using variables of the form dat$var, for example:

fit <- glmer(presence~log(SIZE_strict)*Troph_level+log(HAB500EXCL_strict+1)+
  (1|dataset), family=poisson, data=dat)

来源：https://stackoverflow.com/questions/19431863/lme4-upgrade-produces-error-message-error-in-data-frametmp-i-value