如何在多个维度上聚合data.table

Question

我有一个数据表,我想根据多个字段聚合数据.以下是我的数据的简化示例:

# each record is the number of pages read
# by a student in a given day
pages_per_day <- data.table(
  student_id = c(1,1,1,2,2,2),
  week_of_semester = c(1,1,2,1,2,2),
  pages_read = c(8,6,4,7,8,7)
)

我想根据student_id和week聚合这些数据,以显示在学期的某一周内每位学生阅读的平均页数.我尝试了以下方法:

avg_weekly_pages_read <- grades[,list(
  avg_pages = sum(pages_read) / .N,
  by = c('student_id','week')
)]

这给了我一个包含列的两列数据表:avg_pages,by.

我希望有一张桌子更像:

student_id, week, avg_pages
1,1,7
1,2,4
2,1,7
2,2,7.5

非常感谢任何指导.

Answer 1

你在找

pages_per_day[, .(avg_pages = mean(pages_read)), by = .(student_id, week_of_semester)]
#    student_id week_of_semester avg_pages
# 1:          1                1       7.0
# 2:          1                2       4.0
# 3:          2                1       7.0
# 4:          2                2       7.5

顺便说一句,没有必要重新发明轮子.meanR中有一个功能