Resolved: Deal with occasional numeric(0) inside group_by()

library(tidyverse)
library(lubridate) 

df <- data.frame(
  id = c("a1", "a1", "a1", "b1", "b1", "b1"), 
  month = c("2021-01-01", "2021-03-01", "2021-04-01", "2021-05-01", "2021-06-01", "2021-08-01"), 
  x = c(34,56,76, 12, 13, 14), 
  month_start = c("2021-01-01", "2021-01-01", "2021-01-01", "2021-05-01", "2021-05-01", "2021-05-01")
)

df <- df %>% mutate(month = as.Date(month), month_start = as.Date(month_start))

df %>%
  group_by(id) %>%
  mutate(x_growth_start = x[month == month_start %m+% months(1)],
         x_growth_end = x[month == month_start %m+% months(3)])

df %>%
  group_by(id) %>%
  mutate(x_growth_start = x[match(month_start %m+% months(1), month)],
         x_growth_end   = x[match(month_start %m+% months(3), month)])

#   id    month          x month_start x_growth_start x_growth_end
#   <chr> <date>     <dbl> <date>               <dbl>        <dbl>
# 1 a1    2021-01-01    34 2021-01-01              NA           76
# 2 a1    2021-03-01    56 2021-01-01              NA           76
# 3 a1    2021-04-01    76 2021-01-01              NA           76
# 4 b1    2021-05-01    12 2021-05-01              13           14
# 5 b1    2021-06-01    13 2021-05-01              13           14
# 6 b1    2021-08-01    14 2021-05-01              13           14