左連接沒有正確匹配

Question

我有以下數據：

structure(list(Time = c(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 
1, 1, 1, 1, 1, 1), AgentID = 1:40, State = c(59L, 28L, 84L, 11L, 
5L, 8L, 14L, 71L, 47L, 7L, 84L, 95L, 91L, 92L, 99L, 34L, 70L, 
37L, 55L, 96L, 46L, 38L, 71L, 2L, 61L, 13L, 73L, 26L, 44L, 59L, 
52L, 53L, 42L, 66L, 23L, 11L, 42L, 77L, 38L, 48L), Action = c(-1L, 
-1L, 1L, -1L, 1L, 1L, 1L, -1L, -1L, -1L, -1L, 1L, 1L, 1L, 1L, 
-1L, 1L, 1L, -1L, -1L, 1L, 1L, -1L, 1L, 1L, -1L, -1L, 1L, -1L, 
-1L, 1L, -1L, -1L, 1L, 1L, -1L, -1L, 1L, 1L, 1L), Reward = c(-987L, 
15L, -479L, -485L, -785L, -683L, -1281L, -990L, -886L, -186L, 
-83L, -285L, -886L, -387L, -1087L, -791L, -687L, -988L, -888L, 
-285L, -888L, -690L, -185L, -387L, -789L, -589L, -1089L, -391L, 
-388L, -1193L, 18L, -388L, -989L, -278L, -487L, -988L, -484L, 
-588L, -282L, -790L), ActorLoss = c(-685.0519, 10.296739, -332.56876, 
-339.06058, -543.66394, -471.6095, -890.6096, -685.5919, -615.5027, 
-128.70341, -57.796043, -194.98253, -615.1243, -269.02368, -758.362, 
-548.063, -478.1057, -690.9155, -616.49274, -197.9684, -615.4089, 
-478.90158, -128.84201, -268.25974, -546.7193, -407.66656, -752.44385, 
-270.63773, -268.98254, -825.52856, 12.47267, -268.95764, -684.3579, 
-190.53835, -336.1535, -687.00714, -335.5734, -408.69858, -196.12567, 
-549.0034), CriticLoss = c(346.44806, 3.8356564, 264.62875, 223.61797, 
282.86646, 264.60562, 412.33395, 346.4176, 300.00894, 141.48476, 
100.09644, 223.62798, 331.69186, 200.03246, 360.58798, 316.22833, 
264.60284, 374.19714, 300.0259, 173.24109, 331.69604, 264.6325, 
141.4761, 244.95958, 346.43042, 282.85916, 331.66525, 244.9461, 
244.9759, 374.21875, 4.2367864, 244.98549, 374.17026, 173.27281, 
223.65364, 346.42776, 223.63875, 282.86008, 173.25778, 316.2569
), N = c(40L, 40L, 40L, 40L, 40L, 40L, 40L, 40L, 40L, 40L, 40L, 
40L, 40L, 40L, 40L, 40L, 40L, 40L, 40L, 40L, 40L, 40L, 40L, 40L, 
40L, 40L, 40L, 40L, 40L, 40L, 40L, 40L, 40L, 40L, 40L, 40L, 40L, 
40L, 40L, 40L), SimulationID = c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L), discountFactor = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0)), row.names = c(NA, -40L), class = c("data.table", 
"data.frame"), .internal.selfref = <pointer: 0x55d1abc6fab0>)

我想做的是創建一個新的 data.table，比如dStat ，它將包含Time,SimulationID,N,discountFactor,position 。 前 4 列（ Time,SimulationID,N,discountFactor ）來自上述數據，所有SimulationID,N,discountFactor的position將只是seq(1,100,1) 。

然后我想在dStat中創建一個名為pDensity的新列，這樣pDensity將是上述數據中的 # AgentID 。

我試過了

dStat <- a[, list(position = seq(0,L,1)), by=.(Time,SimulationID,N,discountFactor)]
dStat[a, pDensity:= .N, on=.(position=State,Time,SimulationID,N,discountFactor)]

但是對於所有position s， pDensity為 40，其中至少有一個AgentID與State=position 。

但是 40 是AgentIDs的總數，而不是滿足State=position的 # AgentID的總數。

我在這里做錯了什么？

Answer 1

似乎確實有 40 個代理 ID 滿足state=Position 。
這是合乎邏輯的，因為max(State)=100和position=1:100 ，所以總是有匹配的。

dStat[a,.(Time,SimulationID,N,discountFactor,i.State,x.position,AgentID) , on=.(position=State,Time,SimulationID,N,discountFactor)]
    Time SimulationID  N discountFactor i.State x.position AgentID
 1:    1            1 40              0      59         59       1
 2:    1            1 40              0      28         28       2
 3:    1            1 40              0      84         84       3
 4:    1            1 40              0      11         11       4
 5:    1            1 40              0       5          5       5
 6:    1            1 40              0       8          8       6
 7:    1            1 40              0      14         14       7
 8:    1            1 40              0      71         71       8
 9:    1            1 40              0      47         47       9
10:    1            1 40              0       7          7      10
11:    1            1 40              0      84         84      11
12:    1            1 40              0      95         95      12
13:    1            1 40              0      91         91      13
14:    1            1 40              0      92         92      14
15:    1            1 40              0      99         99      15
16:    1            1 40              0      34         34      16
17:    1            1 40              0      70         70      17
18:    1            1 40              0      37         37      18
19:    1            1 40              0      55         55      19
20:    1            1 40              0      96         96      20
21:    1            1 40              0      46         46      21
22:    1            1 40              0      38         38      22
23:    1            1 40              0      71         71      23
24:    1            1 40              0       2          2      24
25:    1            1 40              0      61         61      25
26:    1            1 40              0      13         13      26
27:    1            1 40              0      73         73      27
28:    1            1 40              0      26         26      28
29:    1            1 40              0      44         44      29
30:    1            1 40              0      59         59      30
31:    1            1 40              0      52         52      31
32:    1            1 40              0      53         53      32
33:    1            1 40              0      42         42      33
34:    1            1 40              0      66         66      34
35:    1            1 40              0      23         23      35
36:    1            1 40              0      11         11      36
37:    1            1 40              0      42         42      37
38:    1            1 40              0      77         77      38
39:    1            1 40              0      38         38      39
40:    1            1 40              0      48         48      40
    Time SimulationID  N discountFactor i.State x.position AgentID

Answer 2

解決了！

dStat <- d[, list(position = seq(0,L,1)), by=.(Time,SimulationID,N,discountFactor)]
dStat[d, pDensity:= .N/N, on=.(position=State,Time,SimulationID,N,discountFactor), by=.(position, Time, SimulationID, N, discountFactor)]

通過將by=.(position, Time, SimulationID, N, discountFactor)添加到連接操作。