我想基于最近一次兩個資料集相結合后，但我不知道如何與指定此data.table。
x 資料集在這里，'y' 資料集如下：
#y data.table
structure(list(date.time = structure(c(1551107760, 1551171420,
1551181500, 1551181560, 1551182400, 1551188100, 1551194700, 1551262320,
1551262380, 1551264420, 1551267960, 1551272400, 1551338580, 1551343140,
1551343380, 1551349800, 1551351600, 1551369000, 1551424680, 1551425520,
1551429000, 1551446880), class = c("POSIXct", "POSIXt"), tzone = "UTC"),
id = c(12L, 3L, 3L, 10L, 10L, 11L, 4L, 10L, 3L, 11L, 4L,
12L, 2L, 10L, 12L, 12L, 12L, 12L, 2L, 12L, 2L, 10L)), row.names = c(NA,
-22L), class = c("data.table", "data.frame"), .internal.selfref = <pointer: 0x000001dc21061ef0>, sorted = "date.time")
如果我指定roll=nearest：
require(data.table)
setDT(x)
setDT(y)
x[, reference.x := date.time] #ref to check join
setkey(y, date.time)
setkey(x, date.time)
#join dataframe
results <- x[y, roll="nearest"]
你會看到，函式作業，因為它應該，但我想加盟是基于最近的一次x發生之后y。在結果中，您將看到 的日期和時間reference.x在 的日期和時間之前y。
> results
date.time class reference.x id
1: 2019-02-25 15:16:00 8.2 14.8 2019-02-25 15:16:00 12 #on or after
2: 2019-02-26 08:57:00 8.2 14.8 2019-02-26 08:57:00 3 #on or after
3: 2019-02-26 11:45:00 8.2 14.8 2019-02-26 11:45:00 3 #on or after
4: 2019-02-26 11:46:00 8.2 14.8 2019-02-26 11:45:00 10 #before (incorrect)
5: 2019-02-26 12:00:00 8.2 14.8 2019-02-26 11:59:00 10
6: 2019-02-26 13:35:00 8.2 14.8 2019-02-26 13:35:00 11
7: 2019-02-26 15:25:00 8.2 14.8 2019-02-26 15:25:00 4
8: 2019-02-27 10:12:00 8.2 14.8 2019-02-27 07:10:00 10
9: 2019-02-27 10:13:00 8.2 14.8 2019-02-27 07:10:00 3
10: 2019-02-27 10:47:00 8.2 14.8 2019-02-27 07:10:00 11
11: 2019-02-27 11:46:00 8.2 14.8 2019-02-27 07:10:00 4
12: 2019-02-27 13:00:00 8.2 14.8 2019-02-27 07:10:00 12
13: 2019-02-28 07:23:00 8.2 14.8 2019-02-28 07:26:00 2
14: 2019-02-28 08:39:00 8.2 14.8 2019-02-28 08:39:00 10
15: 2019-02-28 08:43:00 8.2 14.8 2019-02-28 08:42:00 12
16: 2019-02-28 10:30:00 8.2 14.8 2019-02-28 10:30:00 12
17: 2019-02-28 11:00:00 8.2 14.8 2019-02-28 10:59:00 12
18: 2019-02-28 15:50:00 8.2 14.8 2019-02-28 15:49:00 12
19: 2019-03-01 07:18:00 8.2 14.8 2019-03-01 07:18:00 2
20: 2019-03-01 07:32:00 8.2 14.8 2019-03-01 07:32:00 12
21: 2019-03-01 08:30:00 8.2 14.8 2019-03-01 08:30:00 2
22: 2019-03-01 13:28:00 8.2 14.8 2019-03-01 13:27:00 10
>
structure(list(date.time = structure(c(1551107760, 1551171420,
1551181500, 1551181560, 1551182400, 1551188100, 1551194700, 1551262320,
1551262380, 1551264420, 1551267960, 1551272400, 1551338580, 1551343140,
1551343380, 1551349800, 1551351600, 1551369000, 1551424680, 1551425520,
1551429000, 1551446880), class = c("POSIXct", "POSIXt"), tzone = "UTC"),
class = c("8.2 14.8", "8.2 14.8", "8.2 14.8", "8.2 14.8",
"8.2 14.8", "8.2 14.8", "8.2 14.8", "8.2 14.8", "8.2 14.8",
"8.2 14.8", "8.2 14.8", "8.2 14.8", "8.2 14.8", "8.2 14.8",
"8.2 14.8", "8.2 14.8", "8.2 14.8", "8.2 14.8", "8.2 14.8",
"8.2 14.8", "8.2 14.8", "8.2 14.8"), reference.x = structure(c(1551107760,
1551171420, 1551181500, 1551181500, 1551182340, 1551188100,
1551194700, 1551251400, 1551251400, 1551251400, 1551251400,
1551251400, 1551338760, 1551343140, 1551343320, 1551349800,
1551351540, 1551368940, 1551424680, 1551425520, 1551429000,
1551446820), tzone = "UTC", class = c("POSIXct", "POSIXt"
)), id = c(12L, 3L, 3L, 10L, 10L, 11L, 4L, 10L, 3L, 11L,
4L, 12L, 2L, 10L, 12L, 12L, 12L, 12L, 2L, 12L, 2L, 10L)), sorted = "date.time", class = c("data.table",
"data.frame"), row.names = c(NA, -22L), .internal.selfref = <pointer: 0x000001dc21061ef0>)
任何幫助將不勝感激。
uj5u.com熱心網友回復：
您正在尋找的選項是 roll = -Inf
從檔案 ( ?data.table) ：
-Inf而是向后滾動；即，下一個觀察向后進行（NOCB）。
> results <- x[y, roll=-Inf]
> head(results)
date.time class reference.x id
1: 2019-02-25 15:16:00 8.2 14.8 2019-02-25 15:16:00 12
2: 2019-02-26 08:57:00 8.2 14.8 2019-02-26 08:57:00 3
3: 2019-02-26 11:45:00 8.2 14.8 2019-02-26 11:45:00 3
4: 2019-02-26 11:46:00 8.2 14.8 2019-02-26 11:47:00 10
5: 2019-02-26 12:00:00 8.2 14.8 2019-02-26 12:01:00 10
6: 2019-02-26 13:35:00 8.2 14.8 2019-02-26 13:35:00 11
轉載請註明出處，本文鏈接：https://www.uj5u.com/qiye/375146.html