在 data.table 中分组后查找统计信息答案

【问题标题】：Finding statistics after grouping in data.table在 data.table 中分组后查找统计信息
【发布时间】：2021-12-12 17:49:45
【问题描述】：

我有一个关于 data.table 的小问题。由于我不太擅长，我不太确定如何在 data.table 中做到这一点。

基本上我有 3 列，并希望按前两列（键和日期）分组，然后对于每个键和每个日期，找到第三列（票价）中出现的最大值和最小值

我试过这样做，但它给了我一个错误

flights[, c("max_day", "min_day") := unlist(lapply(gross_fare, findr)), by = c("key", "created_date")]

Error in `[.data.table`(flights, , `:=`(c("max_day", "min_day"), unlist(lapply(gross_fare,  : 
  Supplied 18 items to be assigned to group 1 of size 9 in column 'max_day'. The RHS length must either be 1 (single values are ok) or match the LHS length exactly. If you wish to 'recycle' the RHS please use rep() explicitly to make this intent clear to readers of your code.

findr 是一个只找到最大值和最小值的函数，即

findr <- function(x) {list(max = max(x), min = min(x)}

我已经在 dplyr 中完成了我想做的事情，我会附上代码，但是由于我有数百万行，dplyr 会占用我的内存，所以 data.table 会有所帮助

test <- flights %>%
  select(key, created_date, gross_fare) %>%
  group_by(key, created_date) %>%
  summarise(
            max_day = max(gross_fare),
            min_day = min(gross_fare),
            diff = max_day - min_day) %>%
  arrange(created_date)

如果有人想使用它，我已经把 dput 输出如果有人能帮忙就太好了，谢谢:)

data.table::setDT(structure(list(key = c("LHE_KHI_LHE+KHI_PA-405_15.0_1", "KHI_ISB_KHI+ISB_PK-370_20.0_0", 
"LHE_KHI_LHE+KHI_PK-307_20.0_0", "ISB_KHI_ISB+KHI_PF-124_20.0_1", 
"LHE_KHI_LHE+KHI_PK-307_20.0_0", "LHE_KHI_LHE+KHI_PA-405_15.0_1", 
"KHI_LHE_KHI+LHE_PK-304_20.0_0", "KHI_ISB_KHI+ISB_PA-204_15.0_1", 
"ISB_KHI_ISB+KHI_PA-207_15.0_1", "KHI_ISB_KHI+ISB_PA-200_20.0_1", 
"KHI_LHE_KHI+LHE_PK-304_40.0_0", "ISB_KHI_ISB+KHI_PA-201_35.0_1", 
"ISB_KHI_ISB+KHI_ER-501_20.0_1", "KHI_LHE_KHI+LHE_PF-145_20.0_2", 
"KHI_ISB_KHI+ISB_PA-204_20.0_1", "LHE_KHI_LHE+KHI_PA-401_0.0_0", 
"ISB_KHI_ISB+KHI_PK-309_40.0_0", "KHI_ISB_KHI+ISB_PF-123_20.0_2", 
"ISB_KHI_ISB+KHI_PA-205_15.0_1", "LHE_KHI_LHE+KHI_PF-142_0.0_0", 
"ISB_KHI_ISB+KHI_PA-223_15.0_1", "ISB_KHI_ISB+KHI_PF-126_20.0_2", 
"ISB_KHI_ISB+KHI_PK-309_20.0_0", "KHI_ISB_KHI+ISB_PF-121_20.0_2", 
"ISB_KHI_ISB+KHI_PK-373_20.0_0", "KHI_LHE_KHI+LHE_PF-145_20.0_2", 
"KHI_LHE_KHI+LHE_PA-402_15.0_1", "LHE_KHI_LHE+KHI_PA-407_20.0_1", 
"KHI_ISB_KHI+ISB_PK-308_40.0_0", "KHI_LHE_KHI+LHE_PF-145_20.0_2", 
"LHE_KHI_LHE+KHI_PF-144_0.0_0", "ISB_KHI_ISB+KHI_PK-369_40.0_0", 
"ISB_KHI_ISB+KHI_PF-124_20.0_2", "KHI_ISB_KHI+ISB_PA-204_15.0_1", 
"KHI_ISB_KHI+ISB_PA-200_15.0_1", "ISB_KHI_ISB+KHI_PF-124_20.0_1", 
"KHI_ISB_KHI+ISB_PK-300_20.0_0", "ISB_KHI_ISB+KHI_PF-122_20.0_2", 
"KHI_ISB_KHI+ISB_PK-368_20.0_0", "KHI_ISB_KHI+ISB_PA-204_15.0_1", 
"ISB_KHI_ISB+KHI_ER-503_20.0_1", "ISB_KHI_ISB+KHI_PA-209_15.0_1", 
"KHI_ISB_KHI+ISB_PK-308_40.0_0", "ISB_KHI_ISB+KHI_PF-124_20.0_1", 
"ISB_KHI_ISB+KHI_PK-301_40.0_0", "KHI_LHE_KHI+LHE_PA-408_35.0_1", 
"LHE_KHI_LHE+KHI_PF-144_20.0_2", "KHI_ISB_KHI+ISB_PF-121_20.0_2", 
"KHI_ISB_KHI+ISB_PA-204_35.0_1", "ISB_KHI_ISB+KHI_PK-309_40.0_0", 
"ISB_KHI_ISB+KHI_PA-223_20.0_1", "KHI_ISB_KHI+ISB_PA-206_35.0_1", 
"LHE_KHI_LHE+KHI_PF-142_32.0_1", "LHE_KHI_LHE+KHI_PF-142_20.0_1", 
"KHI_ISB_KHI+ISB_PF-123_20.0_2", "ISB_KHI_ISB+KHI_PA-209_15.0_1", 
"KHI_ISB_KHI+ISB_PA-204_35.0_1", "ISB_KHI_ISB+KHI_PA-201_20.0_1", 
"KHI_ISB_KHI+ISB_PK-368_20.0_0", "ISB_KHI_ISB+KHI_PA-205_20.0_1", 
"KHI_ISB_KHI+ISB_PF-121_20.0_1", "ISB_KHI_ISB+KHI_PF-124_20.0_1", 
"ISB_KHI_ISB+KHI_PA-205_15.0_1", "KHI_LHE_KHI+LHE_PF-145_20.0_2", 
"KHI_LHE_KHI+LHE_PA-406_35.0_1", "KHI_ISB_KHI+ISB_PK-308_20.0_0", 
"LHE_KHI_LHE+KHI_PA-401_20.0_1", "LHE_KHI_LHE+KHI_PA-401_15.0_1", 
"KHI_ISB_KHI+ISB_PA-204_35.0_1", "KHI_LHE_KHI+LHE_PA-406_35.0_1", 
"KHI_ISB_KHI+ISB_PA-206_35.0_1", "KHI_ISB_KHI+ISB_PF-121_20.0_1", 
"ISB_KHI_ISB+KHI_PA-205_20.0_1", "LHE_KHI_LHE+KHI_PF-142_20.0_1", 
"LHE_KHI_LHE+KHI_PF-146_20.0_2", "LHE_KHI_LHE+KHI_PA-401_35.0_1", 
"ISB_KHI_ISB+KHI_PA-209_15.0_1", "ISB_KHI_ISB+KHI_PK-301_40.0_0", 
"ISB_KHI_ISB+KHI_PA-205_35.0_1", "KHI_LHE_KHI+LHE_PA-406_15.0_1", 
"KHI_ISB_KHI+ISB_PF-123_20.0_1", "ISB_KHI_ISB+KHI_PA-201_35.0_1", 
"KHI_ISB_KHI+ISB_PK-300_40.0_0", "KHI_LHE_KHI+LHE_PA-402_35.0_1", 
"ISB_KHI_ISB+KHI_ER-505_20.0_1", "ISB_KHI_ISB+KHI_PF-122_20.0_2", 
"ISB_KHI_ISB+KHI_PA-207_15.0_1", "KHI_LHE_KHI+LHE_PA-404_35.0_1", 
"KHI_ISB_KHI+ISB_PF-123_20.0_1", "ISB_KHI_ISB+KHI_ER-503_20.0_1", 
"ISB_GIL_ISB+GIL_PK-605_20.0_0", "KHI_ISB_KHI+ISB_PF-123_20.0_1", 
"KHI_ISB_KHI+ISB_PA-200_15.0_1", "ISB_KHI_ISB+KHI_PF-122_20.0_2", 
"KHI_LHE_KHI+LHE_PA-404_35.0_1", "ISB_KHI_ISB+KHI_PF-122_20.0_2", 
"PEW_KHI_PEW+KHI_PF-152_20.0_1", "LHE_KHI_LHE+KHI_PK-303_20.0_0", 
"KHI_ISB_KHI+ISB_PA-222_35.0_1", "ISB_KHI_ISB+KHI_PF-124_20.0_1"
), created_date = c("2021-04-20", "2021-05-27", "2021-02-13", 
"2021-08-14", "2021-08-11", "2021-08-21", "2021-01-26", "2021-08-21", 
"2021-05-24", "2021-09-15", "2021-06-05", "2021-07-19", "2021-09-29", 
"2021-07-02", "2021-08-10", "2021-01-04", "2021-07-15", "2021-07-14", 
"2021-08-13", "2021-01-11", "2021-09-13", "2021-09-20", "2021-05-27", 
"2021-02-20", "2021-08-15", "2021-07-27", "2021-08-26", "2021-09-15", 
"2021-08-02", "2021-06-25", "2021-05-15", "2021-08-26", "2021-07-30",
"2021-06-27", "2021-08-07", "2021-03-19", "2021-03-02", "2021-06-06", 
"2021-08-15", "2021-06-27", "2021-09-19", "2021-07-28", "2021-08-09", 
"2021-08-16", "2021-09-09", "2021-06-04", "2021-08-12", "2021-05-15", 
"2021-07-26", "2021-05-27", "2021-08-12", "2021-08-02", "2021-01-26", 
"2021-04-20", "2021-08-26", "2021-08-26", "2021-03-21", "2021-01-09", 
"2021-04-23", "2021-01-04", "2021-08-13", "2021-06-22", "2021-05-31", 
"2021-08-18", "2021-06-16", "2021-08-14", "2021-08-10", "2021-06-16", 
"2021-04-08", "2021-05-20", "2021-06-22", "2021-04-20", "2021-01-05", 
"2021-02-27", "2021-07-07", "2021-03-26", "2021-08-16", "2021-05-01", 
"2021-07-31", "2021-06-14", "2021-06-16", "2021-03-25", "2021-09-14", 
"2021-06-06", "2021-09-02", "2021-08-06", "2021-07-18", "2021-02-28", 
"2021-04-28", "2021-09-19", "2021-08-25", "2021-06-17", "2021-06-07", 
"2021-06-17", "2021-07-07", "2021-08-23", "2021-07-09", "2021-07-19", 
"2021-07-14", "2021-05-21"), gross_fare = c(7796, 7427, 11504, 
6870, 6580, 14945, 8697, 7524, 7124, 6785, 11858, 7524, 11500, 
9525, 6785, 8739, 8200, 13560, 9045, 7400, 7524, 12500, 7458, 
14000, 6570, 9525, 6220, 10545, 8310, 7900, 7820, 8410, 11285, 
19892, 6810, 9800, 11441, 11900, 6570, 13592, 11500, 8300, 20380, 
8525, 7340, 9707, 7870, 10655, 10545, 11798, 14645, 10545, 8650, 
8650, 7870, 12945, 10799, 10227, 6765, 10227, 20120, 11045, 9403, 
7870, 7124, 6570, 6810, 6531, 8605, 7124, 11072, 7390, 10227, 
13435, 10530, 12280, 18945, 11147, 10545, 6531, 6620, 10799, 
18480, 32702, 5606, 13560, 23895, 8027, 9655, 11500, 11990, 6620, 
9403, 7620, 14645, 19105, 9000, 6440, 12645, 8025)), row.names = c(NA, 
-100L), class = c("data.table", "data.frame")))

【问题讨论】：

flights[, ':='(max_day = max(gross_fare),min_day=min(gross_fare),diff= max(gross_fare)-min(gross_fare)), by = c("key" , "created_date")]

标签： r data.table

【解决方案1】：

我想这行代码应该可以完成这项工作：

library(data.table)

flights[,  .(min_day = min(gross_fare), max_day = max(gross_fare), diff = max(gross_fare) - min(gross_fare)), by = .(key, created_date)][]

【讨论】：

这行代码做得很好，如果你不介意解释一下，你能告诉我.( )是什么意思，最后的[]是什么意思
嗨@MSasta。很高兴我能帮助你。 .( ) 部分只是 list( ) 的别名（如果您愿意，可以使用 list( ) 代替）。最后的[ ] 返回/显示结果data.table 的框架。如果你删除这部分（你可以试试看），R 会完成这项工作，但你不会看到结果。

【解决方案2】：

由于函数findr返回一个列表，所以没有必要把事情复杂化：

findr <- function(x) {list(max = max(x), min = min(x))}

flights[, c("max_day", "min_day") := findr(gross_fare), by = list(key, created_date)][]

要同时返回 max 和 min 之间的差异，请使用

findr2 <- function(x) {
  list(max = max(x), min = min(x), diff = diff(range(x)))
}

flights[, c("max_day", "min_day", "diff_day") := findr2(gross_fare), by = list(key, created_date)][]

【讨论】：

我认为汇总会返回比原始数据更少的行；你的第二个街区没有。顺便说一句：如果findr2 的变量名足够好，则可以在没有赋值的情况下执行flights[, findr2(gross_fare), by=...]，并且根据您的第一点，它仍然“正常工作”。