I am unable to reproduce the error in question and cannot attach my full data. Any help understanding the error received would be appreciated.

After imputation, I make several changes to the imputed datasets and would like to convert the transformed data to a mids object for modeling. However, I get the following error:

Error in `[<-.data.frame`(`*tmp*`, j, value = list(variable = c(0.9998,  : 
  replacement element 1 has 78 rows, need 86

Below is a toy example of steps I take:

library(mice)
library(dplyr)

df <- data.frame(
  id = rep(1:3, 3),
  a = sample(c(NA, 1, 0), size=9, replace = T),
  b = sample(c(NA, 1:10), size=9, replace = T)
)

# perform imputation
imp <- mice(df)

# get complete data + transform
comp <- complete(imp, action="long", include = T) %>%
  filter(id %in% 1:2) %>%
  rowwise() %>%
  mutate(add = sum(a,b, na.rm = T)) %>%
  dplyr::ungroup()

# convert to mids object - error encountered here
to_mids <- as.mids(comp)