【R|知其然】anova_test 报错 Can‘t subset columns that don‘t exist.

【知其然】意为知其然但不知其所以然，记录一些莫名其妙就解决了的bug，但并不知道为什么这样能解决

在Rstudio中想对数据进行方差分析时，出现了报错。

方差分析代码：

aov_beha <-behav %>% 
  anova_test(
    data = ., dv = RT, wid = sub, 
    within = cond
  )

get_anova_table(aov_beha)

报错内容

ERROR: Problem with `mutate()` input `data`.
x Can't subset columns that don't exist.
x Column `sub` doesn't exist.
i Input `data` is `map(.data$data, .f, ...)`.

最初我以为是数据类型的问题，转来转去转了几次后，依然还是报错

> head(behav)
# A tibble: 6 x 4
# Groups:   sub [2]
    sub  cond    RT   ACC
     
1     1     1 0.311 1    
2     1     2 0.256 1    
3     1     3 0.317 1    
4     1     4 0.305 0.967
5     2     1 0.328 1    
6     2     2 0.339 0.967

最后才发现，原来是我之前对数据做了group_by处理。解决办法是，加上ungroup就行了

behav <- od_beha %>% 
  group_by(sub,cond) %>% 
  summarise(RT = mean(RT), ACC= mean(ACC)) %>% 
   ungroup()

ref: r - Error: Can't subset columns that don't exist - Stack Overflow