Project Spatial Tables Without Changing Column Metadata
Source:R/project-spatial-table.R
ProjectSpatialTable.RdSelects explicit rows and columns from a plain data frame or data.table.
Row order and repetitions follow rows; column order follows columns.
Selection uses a copied data.table inside the package namespace, preserving
column storage types, classes and attributes according to data.table subset
semantics. The result is a plain data frame with reset row names.
No eligibility, missing-value filtering, sorting or numerical transformation
is inferred. The supplied input is not modified.
Arguments
- data
A plain data.frame or data.table with distinct, nonempty column names. Other data-frame subclasses are not accepted.
- rows
A dimensionless, unclassed integer vector of positive row indices no greater than the number of input rows. Missing indices are not allowed.
integer()selects zero rows; repeated and shuffled indices are retained.- columns
A dimensionless, unclassed character vector of distinct, nonempty existing column names, in the requested order.
Value
A plain data.frame with the requested rows and columns and reset row names. A one-column selection remains a data frame; an empty row selection retains typed columns and their metadata. Invalid inputs are rejected rather than coerced, recycled or partially matched.
Examples
d <- data.frame(id = letters[1:3], weight = c(0.5, 1.5, 2.5))
ProjectSpatialTable(d, c(3L, 1L, 3L), c("weight", "id"))
#> weight id
#> 1 2.5 c
#> 2 0.5 a
#> 3 2.5 c
ProjectSpatialTable(d, integer(), "weight")
#> [1] weight
#> <0 rows> (or 0-length row.names)