根据标签对数据集进行分组

Question

我有一个来自游戏应用程序的数据集。 我想将我的应用程序数据集划分为多个会话。新的会话以appstart或appresume开始。它大多以appclose结尾，但并非总是如此。

记录结构如下：

appdata <- data.frame(userid = c(1,1,1,1,1,1,1), activity = c("appstart","appclose","appresume","run","appstart", "run", "run"),points = c(263,275,278,321, 322, 412, 498))

appdata
  userid  activity points
1      1  appstart    263
2      1  appclose    275
3      1 appresume    278
4      1       run    321
5      1  appstart    322
6      1       run    412
7      1       run    498

我希望它看起来像这样：

appdata_new <- data.frame(userid = c(1,1,1), session = c(1,2,3), mean_points = c(269, 299.5, 410.67))

appdata_new
  userid session mean_points
1      1       1      269.00
2      1       2      299.50
3      1       3      410.67

根据用户的评论，我尝试了以下操作，但不适用于我的数据集：

appdata %>% group_by(userid, session = cumsum(lag(activity == "appstart", default = FALSE) | cumsum(lag(activity == "appresume", default = FALSE))) %>% summarise(mean_points = mean(points))