Lean = TRUE not working with interacted IV

[adamaltmejd]

Hey! Just found a bug running interacted IV, following up on this discussion: #99

data(base_did)
base_did$takeup <- rbinom(nrow(base_did), 1, 0.5)
# est_did = feols(y ~ x1 | id + period | i(takeup, period, 5) ~ i(treat, period, 5), base_did) #works
est_did = feols(y ~ x1 | id + period | i(takeup, period, 5) ~ i(treat, period, 5), base_did, lean = TRUE)

Error in `colnames<-`(`*tmp*`, value = paste0("fit_", iv_lhs_names)) : 
  attempt to set 'colnames' on an object with less than two dimensions

[lrberge]

Thanks! It actually occurred whenever the IV estimation contained FEs!
Big thanks for the reproducible example => quick to fix!
(Confession: sometimes I really feel like Sisyphus!)