Revision 38aa5c07
Added by Jim Regetz over 12 years ago
- ID 38aa5c074cc9fec577465c93f5a25924b8cd5a3c
| climate/extra/ghcn-to-sqlite.R | ||
|---|---|---|
| 79 | 79 |
} |
| 80 | 80 |
} |
| 81 | 81 |
|
| 82 |
# split data columnwise by day, then recombine into long format; note |
|
| 83 |
# that the indexing here is hard-coded to work for the *.dly files, and |
|
| 84 |
# simply assumes that they are all consistent |
|
| 85 |
wideToLong <- function(dat, days) {
|
|
| 86 |
daily.data <- lapply(seq_along(days), function(i) {
|
|
| 87 |
dat <- data.frame(dat[1:4], day=i, dat[days[[i]]]) |
|
| 88 |
dat$srcrowid <- seq(nrow(dat)) |
|
| 89 |
names(dat) <- 1:ncol(dat) |
|
| 90 |
dat |
|
| 91 |
}) |
|
| 92 |
do.call("rbind", daily.data)
|
|
| 93 |
} |
|
| 94 |
|
|
| 95 |
NUM.WIDE.COLS <- 4 + 4*31 |
|
| 96 |
DAYS <- lapply(seq(from=5, to=NUM.WIDE.COLS, by=4), function(i) i:(i+3)) |
|
| 82 | 97 |
dailies <- list.files(pattern="*.dly") |
| 83 | 98 |
for (file in dailies) {
|
| 84 |
x <- loadAsCSV(file, c("TMAX", "TMIN"))
|
|
| 85 |
if (is.null(x)) {
|
|
| 86 |
message("no rows imported for ", file)
|
|
| 87 |
} else {
|
|
| 88 |
long <- reshape(x, direction="long", |
|
| 89 |
varying=matrix(5:ncol(x), nrow=4)) |
|
| 99 |
dly <- loadAsCSV(file, c("TMAX", "TMIN"))
|
|
| 100 |
if (!is.null(dly)) {
|
|
| 101 |
long <- wideToLong(dly, DAYS) |
|
| 90 | 102 |
ghcn_bulk_insert(db, sql, long) |
| 103 |
} else {
|
|
| 104 |
message("no rows imported for ", file)
|
|
| 91 | 105 |
} |
| 92 | 106 |
} |
Also available in: Unified diff
replaced reshape with faster manual split-rbind approach