在R中对大规模数据集实现动态分组基础计算的技术问询

数据集背景 #

我有一个名为df的数据集，包含数万条观测记录，分类变量有100余种。数据集记录了不同个体（id）在指定年份、以指定价格将不同类型患者送往不同地点的信息，示例数据如下：

year <- c(2010, 2010, 2010, 2010, 2011, 2011, 2011, 2010, 2011)
id <- c("A", "A" , "A" , "A" , "A" , "A" , "A", "B", "B")
type <- c("kid", "kid", "adult", "kid", "kid", "dog", "cat", "kid", "kid")
place <- c("hosp", "hosp", "house", "hosp", "hosp", "hosp", "house", "hosp", "hosp")
price <- c(2, 3, 6, 5, 1, 2, 3, 4, 5)

df <- data.frame(year, id, type, place, price)

具体处理需求 #

需按id-year分组对df进行基础统计计算，具体要求：

• 按患者类型生成经验变量：取值为该id拥有该类型的年份数量；
• 按地点生成经验变量：取值为该id拥有该地点的年份数量；
• 计算每个id在当年的平均就诊价格；
• 生成标记变量：该id是否在次年（t+1）出现，取值为0（否）或1（是）。

期望输出 #

最终期望得到类似df_new的数据集：

year <- c("2010", "2011", "2010", "2011")
id <- c("A", "A", "B", "B")
exp_type_kid <- c(1, 2, 1, 2)
exp_type_adult <- c(1, 1, 0, 0)
exp_type_dog <- c(0, 1, 0, 0)
exp_type_cat <- c(0, 1, 0, 0)
exp_place_hosp <- c(1, 2, 1, 2)
exp_place_house <- c(1, 2, 0, 0)
avg_price <- c(4, 2, 4, 5)
id_repeat_next_year <- c(1, 0, 1, 0)

df_new <- data.frame(year, id, exp_type_kid, exp_type_adult, exp_type_dog, exp_type_cat,
                     exp_place_hosp, exp_place_house, avg_price, id_repeat_next_year)

补充说明 #

数据集可能包含更多不连续年份，示例如下：

year <- c(2010, 2010, 2010, 2010, 2011, 2011, 2011, 2009, 2010, 2015, 2017)
id <- c("A", "A" , "A" , "A" , "A" , "A" , "A", "B", "B", "B", "B")
type <- c("kid", "kid", "adult", "kid", "kid", "dog", "cat", "kid", "kid", "kid", "kid")
place <- c("hosp", "hosp", "house", "hosp", "hosp", "hosp", "house", "hosp", "hosp", "hosp", "hosp")
price <- c(2, 3, 6, 5, 1, 2, 3, 4, 4, 4, 4)

df <- data.frame(year, id, type, place, price)

内容的提问来源于stack exchange，提问作者vog