setcolorder gains before= and after=

NEWS.md

@@ -137,6 +137,8 @@
 
 24. `DT[, head(.SD,n), by=grp]` and `tail` are now optimized when `n>1`, [#5060](https://github.com/Rdatatable/data.table/issues/5060) [#523](https://github.com/Rdatatable/data.table/issues/523#issuecomment-162934391). `n==1` was already optimized. Thanks to Jan Gorecki and Michael Young for requesting, and Benjamin Schwendinger for the PR.
 
+25 `setcolorder()` gains `before=` and `after=`, [#4385](https://github.com/Rdatatable/data.table/issues/4358). Thanks to Matthias Gomolka for the request, and both Benjamin Schwendinger and Xianghui Dong for implementing.
+
 ## BUG FIXES
 
 1. `by=.EACHI` when `i` is keyed but `on=` different columns than `i`'s key could create an invalidly keyed result, [#4603](https://github.com/Rdatatable/data.table/issues/4603) [#4911](https://github.com/Rdatatable/data.table/issues/4911). Thanks to @myoung3 and @adamaltmejd for reporting, and @ColeMiller1 for the PR. An invalid key is where a `data.table` is marked as sorted by the key columns but the data is not sorted by those columns, leading to incorrect results from subsequent queries.

R/data.table.R

@@ -2658,15 +2658,21 @@ setnames = function(x,old,new,skip_absent=FALSE) {
   invisible(x)
 }
 
-setcolorder = function(x, neworder=key(x))
+setcolorder = function(x, neworder=key(x), before=NULL, after=NULL)  # before/after #4358
 {
   if (is.character(neworder) && anyDuplicated(names(x)))
     stopf("x has some duplicated column name(s): %s. Please remove or rename the duplicate(s) and try again.", brackify(unique(names(x)[duplicated(names(x))])))
-  # if (!is.data.table(x)) stopf("x is not a data.table")
+  if (!is.null(before) && !is.null(after))
+    stopf("Provide either before= or after= but not both")
+  if (length(before)>1 || length(after)>1)
+    stopf("before=/after= accept a single column name or number, not more than one")
   neworder = colnamesInt(x, neworder, check_dups=FALSE)  # dups are now checked inside Csetcolorder below
+  if (length(before))
+    neworder = c(setdiff(seq_len(colnamesInt(x, before) - 1L), neworder), neworder)
+  if (length(after))
+    neworder = c(setdiff(seq_len(colnamesInt(x, after)), neworder), neworder)
   if (length(neworder) != length(x)) {
-    #if shorter than length(x), pad by the missing
-    #  elements (checks below will catch other mistakes)
+    # pad by the missing elements (checks inside Csetcolorder catch other mistakes)
     neworder = c(neworder, setdiff(seq_along(x), neworder))
   }
   .Call(Csetcolorder, x, neworder)

inst/tests/tests.Rraw

@@ -1529,14 +1529,22 @@ DT = data.table(a=1:2,b=3:4,c=5:6)
 test(495.1, setcolorder(DT,c(2,1,3)), data.table(b=3:4,a=1:2,c=5:6))
 test(495.2, setcolorder(DT,c(2,1,3)), data.table(a=1:2,b=3:4,c=5:6))
 test(496, setcolorder(DT,c("c","a","b")), data.table(c=5:6,a=1:2,b=3:4))
-test(497, setcolorder(DT,c("d","a","b")), error="specify non existing column*.*d")
+test(497.01, setcolorder(DT,c("d","a","b")), error="specify non existing column*.*d")
 DT = data.table(a = 1:3, b = 2:4, c = 3:5)
-test(498.1, names(setcolorder(DT, "b")), c("b", "a", "c"))
-test(498.2, names(setcolorder(DT, c(2, 3))), c("a", "c", "b"))
-test(498.3, setcolorder(DT, 1:4), error = "specify non existing column*.*4")
+test(497.02, names(setcolorder(DT, "b")), c("b", "a", "c"))
+test(497.03, names(setcolorder(DT, c(2, 3))), c("a", "c", "b"))
+test(497.04, setcolorder(DT, 1:4), error = "specify non existing column*.*4")
 # Test where neworder=NULL, thus ordered by key and index columns
 DT = data.table(a = 1:3, b = 2:4, c = 3:5, d = 4:6, key="b")
-test(498.4, names(setcolorder(DT)), c("b", "a", "c", "d"))
+test(497.05, names(setcolorder(DT)), c("b", "a", "c", "d"))
+# new arguments before= and after=, #4358
+DT = data.table(a=1, b=2, c=3)
+test(498.01, setcolorder(DT, "a", after="c"), data.table(b=2, c=3, a=1))
+test(498.02, setcolorder(DT, "a", before="b"), data.table(a=1, b=2, c=3))
+test(498.03, setcolorder(DT, 1, after=3), data.table(b=2, c=3, a=1))
+test(498.04, setcolorder(DT, 3, before=1), data.table(a=1, b=2, c=3))
+test(498.05, setcolorder(DT, 1, before=1, after=1), error="Provide either before= or after= but not both")
+test(498.06, setcolorder(DT, 1, before=1:2), error="before=/after= accept a single column name or number, not more than one")
 
 # test first group listens to nomatch when j uses join inherited scope.
 x <- data.table(x=c(1,3,8),x1=10:12, key="x")

man/setcolorder.Rd

@@ -9,11 +9,12 @@
 }
 
 \usage{
-setcolorder(x, neworder=key(x))
+setcolorder(x, neworder=key(x), before=NULL, after=NULL)
 }
 \arguments{
   \item{x}{ A \code{data.table}. }
   \item{neworder}{ Character vector of the new column name ordering. May also be column numbers. If \code{length(neworder) < length(x)}, the specified columns are moved in order to the "front" of \code{x}. By default, \code{setcolorder} without a specified \code{neworder} moves the key columns in order to the "front" of \code{x}. }
+  \item{before, after}{ If one of them (not both) was provided with a column name or number, \code{neworder} will be inserted before or after that column. }
 }
 \details{
   To reorder \code{data.table} columns, the idiomatic way is to use \code{setcolorder(x, neworder)}, instead of doing \code{x <- x[, neworder, with=FALSE]}. This is because the latter makes an entire copy of the \code{data.table}, which maybe unnecessary in most situations. \code{setcolorder} also allows column numbers instead of names for \code{neworder} argument, although we recommend using names as a good programming practice.