Newer
Older
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
#'Function to Split Dimension
#'
#'@author Nuria Perez-Zanon, \email{nuria.perez@bsc.es}
#'
#'@description This function split a dimension in two. The user can select the dimension to split and provide indices indicating how to split that dimension or dates and the frequency expected (monthly or by day, moth and year). The user can also provide a numeric frequency indicating the length of each division.
#'
#'@param data a 's2dv_cube' object
#'@param dimname a character string indicating the name of the dimension to split
#'@param indices a vector of numeric indices or dates
#'@param freq a character string indicating the frequency given parameter indices as dates.
#'
#'@import abind
#'@import s2dverification
#'@examples
#'
#'data <- 1 : 20
#'dim(data) <- c(time = 10, lat = 2)
#'data <-list(data = data)
#'class(data) <- 's2dv_cube'
#'indices <- c(rep(1,5), rep(2,5))
#'new_data <- CST_SplitDim(data, indices = indices)
#'time <- c(seq(ISOdate(1903, 1, 1), ISOdate(1903, 1, 4), "days"),
#' seq(ISOdate(1903, 2, 1), ISOdate(1903, 2, 4), "days"),
#' seq(ISOdate(1904, 1, 1), ISOdate(1904, 1, 2), "days"))
#'data <- list(data = data$data, Dates = time)
#'class(data) <- 's2dv_cube'
#'new_data <- CST_SplitDim(data, indices = time)
#'dim(new_data$data)
#'new_data <- CST_SplitDim(data, indices = time, freq = 'day')
#'dim(new_data$data)
#'new_data <- CST_SplitDim(data, indices = time, freq = 'month')
#'dim(new_data$data)
#'new_data <- CST_SplitDim(data, indices = time, freq = 'year')
#'dim(new_data$data)
#'@export
CST_SplitDim <- function(data, split_dim = 'time', indices = NULL,
freq = 'monthly') {
if (!inherits(data, 's2dv_cube')) {
stop("Parameter 'data' must be of the class 's2dv_cube', ",
"as output by CSTools::CST_Load.")
}
if (is.null(indices)) {
indices = data$Dates[[1]]
}
data$data <- SplitDim(data$data, split_dim = split_dim, indices = indices,
freq = freq)
return(data)
}
#'Function to Split Dimension
#'
#'@author Nuria Perez-Zanon, \email{nuria.perez@bsc.es}
#'
#'@description This function split a dimension in two. The user can select the dimension to split and provide indices indicating how to split that dimension or dates and the frequency expected (monthly or by day, moth and year). The user can also provide a numeric frequency indicating the length of each division.
#'
#'@param data an n-dimensional array with named dimensions
#'@param dimname a character string indicating the name of the dimension to split
#'@param indices a vector of numeric indices or dates
#'@param freq a character string indicating the frequency given parameter indices as dates.
#'
#'@import abind
#'@import s2dverification
#'@examples
#'
#'data <- 1 : 20
#'dim(data) <- c(time = 10, lat = 2)
#'indices <- c(rep(1,5), rep(2,5))
#'new_data <- SplitDim(data, indices = indices)
#'time <- c(seq(ISOdate(1903, 1, 1), ISOdate(1903, 1, 4), "days"),
#' seq(ISOdate(1903, 2, 1), ISOdate(1903, 2, 4), "days"),
#' seq(ISOdate(1904, 1, 1), ISOdate(1904, 1, 2), "days"))
#'new_data <- SplitDim(data, indices = time)
#'new_data <- SplitDim(data, indices = time, freq = 'day')
#'new_data <- SplitDim(data, indices = time, freq = 'month')
#'new_data <- SplitDim(data, indices = time, freq = 'year')
#'@export
SplitDim <- function(data, split_dim = 'time', indices, freq = 'monthly') {
# check data
if (is.null(data)) {
stop("Parameter 'data' cannot be NULL.")
}
if (is.null(dim(data))) {
dim(data) = c(time = length(data))
}
if (is.null(names(dim(data)))) {
stop("Parameter 'data' must have dimension names.")
}
dims <- dim(data)
# check split_dim
if (!is.character(split_dim)) {
stop("Parameter 'split_dim' must be a character.")
}
if (length(split_dim) > 1) {
split_dim <- split_dim[1]
warning("Parameter 'split_dim' has length greater than ",
"one and only the first element will be used.")
}
if (!any(names(dims) %in% split_dim)) {
stop("Parameter 'split_dims' must be one of the dimension ",
"names in parameter 'data'.")
}
pos_split <- which(names(dims) == split_dim)
# check indices and freq
if (is.null(indices)) {
if (!is.numeric(freq)) {
stop("Parameter 'freq' must be a integer number indicating ",
" the length of each chunk.")
} else {
if (!((dims[pos_split] / freq) %% 1 == 0)) {
stop("Parameter 'freq' must be proportional to the ",
"length of the 'split_dim' in parameter 'data'.")
}
indices <- rep(1 : (dims[pos_split] / freq), freq)
indices <- sort(indices)
}
} else if (is.numeric(indices)) {
if (!is.null(freq)) {
if (freq != 'monthly') {
warning("Parameter 'freq' is not being used since ",
"parameter 'indices' is numeric.")
}
}
} else {
# Indices should be Dates and freq character
if (!is.character(freq)) {
stop("Parameter 'freq' must be a character indicating ",
"how to divide the dates provided in parameter 'indices'",
", 'monthly', 'anually' or 'daily'.")
}
if (!(any(class(indices) %in% c('POSIXct')))) {
indices <- try( {
if (is.character(indices)) {
as.POSIXct(indices)
} else {
as.POSIXct(indices)
}
})
if ('try-error' %in% class(indices) |
sum(is.na(indices)) == length(indices)) {
stop("Dates provided in parameter 'indices' must be of class",
" 'POSIXct' or convertable to 'POSIXct'.")
}
}
}
if (length(indices) != dims[pos_split]) {
stop("Parameter 'indices' has different length of parameter ",
"data in the dimension supplied in 'split_dim'.")
}
# check indices as dates:
if (!is.numeric(indices)) {
if (freq == 'day') {
indices <- as.numeric(strftime(indices, format = "%d"))
} else if (freq == 'month') {
indices <- as.numeric(strftime(indices, format = "%m"))
} else if (freq == 'year') {
indices <- as.numeric(strftime(indices, format = "%Y"))
} else if (freq == 'monthly' ) {
indices <- as.numeric(strftime(indices, format = "%m%Y"))
} else {
stop("Parameter 'freq' must be numeric or a character: ",
"by 'day', 'month', 'year' or 'monthly' (for distingible month).")
}
}
repited <- unique(indices)
max_times <- max(unlist(lapply(repited,
function(x){sum(indices == x)})))
data <- lapply(repited, function(x) {rebuild(x, data, along = split_dim,
indices = indices, max_times)})
data <- abind(data, along = length(dims) + 1)
if (is.character(freq)) {
names(dim(data)) <- c(names(dims), freq)
} else {
names(dim(data)) <- c(names(dims), 'index')
}
return(data)
}
rebuild <- function(x, data, along, indices, max_times) {
a <- Subset(data, along = along, indices = which(indices == x))
pos_dim <- which(names(dim(a)) == along)
if (dim(a)[pos_dim] != max_times) {
adding <- max_times - dim(a)[pos_dim]
new_dims <- dim(a)
new_dims[pos_dim] <- adding
extra <- array(NA, dim = new_dims)
a <- abind(a, extra, along = pos_dim)
names(dim(a)) <- names(dim(data))
}
return(a)
}