First I'd like to thank the creators and maintainers of the R package data.table.
I have a simple question on the 'shift' feature in data.table.
Is there a way not to enter an hardcoded offset but rather a variable offset stored e.g. in a column? I.e. on each row the offset could be different.
I didn't see anything in this direction in the current documentation.
Thank you very much in advance for your kind help.
The code below doesn't work (a vector is displayed instead):
DT[, Observation_Annual_Offset := shift(Observation, n = ifelse (Frequency == "A (Annual)", 1, ifelse (Frequency == "H (Half-yearly)", 2, ifelse (Frequency == "Q (Quarterly)", 4, ifelse(Frequency == "M (Monthly)", 12, 0)))), fill = NA, type = "lead")]
It would be better if you store the frequency in a vector to be referenced rather than using multiple ifelse:
freq <- c("A (Annual)"=1, "H (Half-yearly)"=2, "Q (Quarterly)"=4, "M (Monthly)"=12)
DT <- data.table(Observation=rep(1:25, length(freq)), Offset=rep(names(freq), each=25))
DT[, Observation_Annual_Offset := shift(Observation, -freq[Offset[1L]]), Offset]
data:
Observation Offset Observation_Annual_Offset
1: 1 A (Annual) 2
2: 2 A (Annual) 3
3: 3 A (Annual) 4
4: 4 A (Annual) 5
5: 5 A (Annual) 6
6: 6 A (Annual) 7
7: 7 A (Annual) 8
8: 8 A (Annual) 9
9: 9 A (Annual) 10
10: 10 A (Annual) 11
11: 11 A (Annual) 12
12: 12 A (Annual) 13
13: 13 A (Annual) 14
14: 14 A (Annual) 15
15: 15 A (Annual) 16
16: 16 A (Annual) 17
17: 17 A (Annual) 18
18: 18 A (Annual) 19
19: 19 A (Annual) 20
20: 20 A (Annual) 21
21: 21 A (Annual) 22
22: 22 A (Annual) 23
23: 23 A (Annual) 24
24: 24 A (Annual) 25
25: 25 A (Annual) NA
26: 1 H (Half-yearly) 3
27: 2 H (Half-yearly) 4
28: 3 H (Half-yearly) 5
29: 4 H (Half-yearly) 6
30: 5 H (Half-yearly) 7
31: 6 H (Half-yearly) 8
32: 7 H (Half-yearly) 9
33: 8 H (Half-yearly) 10
34: 9 H (Half-yearly) 11
35: 10 H (Half-yearly) 12
36: 11 H (Half-yearly) 13
37: 12 H (Half-yearly) 14
38: 13 H (Half-yearly) 15
39: 14 H (Half-yearly) 16
40: 15 H (Half-yearly) 17
41: 16 H (Half-yearly) 18
42: 17 H (Half-yearly) 19
43: 18 H (Half-yearly) 20
44: 19 H (Half-yearly) 21
45: 20 H (Half-yearly) 22
46: 21 H (Half-yearly) 23
47: 22 H (Half-yearly) 24
48: 23 H (Half-yearly) 25
49: 24 H (Half-yearly) NA
50: 25 H (Half-yearly) NA
51: 1 Q (Quarterly) 5
52: 2 Q (Quarterly) 6
53: 3 Q (Quarterly) 7
54: 4 Q (Quarterly) 8
55: 5 Q (Quarterly) 9
56: 6 Q (Quarterly) 10
57: 7 Q (Quarterly) 11
58: 8 Q (Quarterly) 12
59: 9 Q (Quarterly) 13
60: 10 Q (Quarterly) 14
61: 11 Q (Quarterly) 15
62: 12 Q (Quarterly) 16
63: 13 Q (Quarterly) 17
64: 14 Q (Quarterly) 18
65: 15 Q (Quarterly) 19
66: 16 Q (Quarterly) 20
67: 17 Q (Quarterly) 21
68: 18 Q (Quarterly) 22
69: 19 Q (Quarterly) 23
70: 20 Q (Quarterly) 24
71: 21 Q (Quarterly) 25
72: 22 Q (Quarterly) NA
73: 23 Q (Quarterly) NA
74: 24 Q (Quarterly) NA
75: 25 Q (Quarterly) NA
76: 1 M (Monthly) 13
77: 2 M (Monthly) 14
78: 3 M (Monthly) 15
79: 4 M (Monthly) 16
80: 5 M (Monthly) 17
81: 6 M (Monthly) 18
82: 7 M (Monthly) 19
83: 8 M (Monthly) 20
84: 9 M (Monthly) 21
85: 10 M (Monthly) 22
86: 11 M (Monthly) 23
87: 12 M (Monthly) 24
88: 13 M (Monthly) 25
89: 14 M (Monthly) NA
90: 15 M (Monthly) NA
91: 16 M (Monthly) NA
92: 17 M (Monthly) NA
93: 18 M (Monthly) NA
94: 19 M (Monthly) NA
95: 20 M (Monthly) NA
96: 21 M (Monthly) NA
97: 22 M (Monthly) NA
98: 23 M (Monthly) NA
99: 24 M (Monthly) NA
100: 25 M (Monthly) NA
Observation Offset Observation_Annual_Offset