Resample csquares objects to higher or lower resolutions.
resample_csquares(x, method = "target", ..., resolution, magnitude = 1L)A csquares object to be resampled to a different resolution
Method for determining the resolution of the resulting csquares.
Should be one of "target", "min", "max", "up", or "down".
"target" will resample x to the level specified with resolution
When x inherits the stars class and the resulting object has
a lower resolution than x, the dots are passed on to dplyr::summarise().
This allows you to summarise columns to the lower resolution.
Resolution (in WGS84 degrees) to be used for creating c-squares codes. As per c-square specifications, the resolution should be 10 or less, yet greater than 0. It should be a tenfold of 1 or 5. Valid resolutions are therefore: 10, 5, 1, 0.5, 0.1, etc.
When method == "up" or "down", this parameter specifies
the number of steps to increase or decrease the resolution. Should be a positive
integer.
A csquares object based on x
csq <- as_csquares(c("1000", "5000:2|5000:100", "3000:100:100"))
csq_df <- as_csquares(data.frame(csq = csq, foobar = letters[1:3]), csquares = "csq")
## Resample csquares based on the one with the lowest resolution:
resample_csquares(csq, "min")
#> csquares [1:3] N 0 E 0 (10°), S 0 W 0 (10°), S 0 E 0 (10°)
## Resample csquares to a specific resolution
resample_csquares(csq, "target", resolution = 5)
#> csquares [1:3] 4 squares, 2 squares, S 0 E 0 (5°)
## Same, but applied to a csquares object inheriting from a data.frame
resample_csquares(csq_df, "target", resolution = 5)
#> csq foobar
#> 1 4 squares a
#> 2 2 squares b
#> 3 S 0 E 0 (5°) c
## Same, but applied to a csquares object inheriting the `sf` class
## Note that the geometry is updated based on the resampled csquares
if (requireNamespace("sf")) {
library(sf)
csq_sf <- st_as_sf(csq_df)
resample_csquares(csq_sf, "target", resolution = 5)
}
#> Simple feature collection with 3 features and 2 fields
#> Geometry type: POLYGON
#> Dimension: XY
#> Bounding box: xmin: -10 ymin: -5 xmax: 10 ymax: 10
#> Geodetic CRS: WGS 84
#> csq foobar geom
#> 1 4 squares a POLYGON ((0 5, 0 10, 5 10, ...
#> 2 2 squares b POLYGON ((-10 -5, -10 0, -5...
#> 3 S 0 E 0 (5°) c POLYGON ((0 0, 5 0, 5 -5, 0...
## Resample csquares one step down.
resample_csquares(csq, "down")
#> csquares [1:3] N 0 E 0 (10°), 2 squares, S 0 E 0 (0.5°)
resample_csquares(csq_df, "down")
#> csq foobar
#> 1 N 0 E 0 (10°) a
#> 2 2 squares b
#> 3 S 0 E 0 (0.5°) c
if (requireNamespace(c("dplyr", "stars"))) {
## Csquares objects can inherit from the stars class as well.
## These too can be resampled. But additional columns need
## to be summarised when the resulting resolution is lower
## than the original:
g <-
sf::st_bbox(c(xmin = 4.0, xmax = 6.5, ymin = 52.5, ymax = 53), crs = 4326) |>
new_csquares(resolution = 0.1) |>
## add a column with some random positive numbers:
dplyr::mutate(random = .data$csquares |> length() |> rnorm() |> exp())
## Resample stars object to lower resolution
g_sum <- resample_csquares(g, resolution = 10, random = sum(random, na.rm = TRUE))
## And back to a higher resolution (note that you have lost information as it was summarised
## in the previous step)
resample_csquares(g_sum, "up", random = sum(random, na.rm = TRUE))
}
#> stars object with 2 dimensions and 2 attributes
#> attribute(s):
#> csquares random
#> Length:4 Min. :224.3
#> Class :character 1st Qu.:224.3
#> Mode :character Median :224.3
#> Mean :224.3
#> 3rd Qu.:224.3
#> Max. :224.3
#> dimension(s):
#> from to offset delta refsys x/y
#> x 1 2 0 5 WGS 84 [x]
#> y 1 2 60 -5 WGS 84 [y]