根据组的第一行变换组中的每一行-Mutateeachrowingroupaccordingtothefirstrowofthegroup

4  

We can do

我们可以做的

 df %>%
  group_by(grp) %>%
  arrange(idx) %>%
  mutate(val = val/sum((row_number() == 1)*val))
# A tibble: 10 x 3
# Groups:   grp [2]
#     grp   idx   val
#     
# 1     1     1 1.000
# 2     2     1 1.000
# 3     1     2 1.500
# 4     2     2 0.625
# 5     1     3 0.250
# 6     2     3 0.375
# 7     1     4 1.750
# 8     2     4 1.125
# 9     1     5 0.500
#10     2     5 0.125

If we need to divide by the first 'val' observation, just do val[1L]

如果我们需要除以第一个'val'观察，只需要做val [1L]

df %>%
   group_by(grp) %>%
   arrange(idx) %>%
   mtuate(val = val/val[1L])