Shorten column names for merge on unnamed objects

The column names of a merge() result can be monstrous for raw, unnamed
numeric objects. Something similar to dput() for the entire column is
used, but zoo deparses the dput() output and limits the deparse length
to ~60 characters.

Use a modified version of makeNames() (defined in merge.zoo()) to make
the column names. Major differences are:
  - Use substitute(alist(...)) to avoid evaluating '...'
  - Use deparse(x, nlines = 1L) to deparse only the first expression

NOTE: this may break existing code that relies on behavior for integer
objects.

For example, current behavior is:

  x <- .xts(1:5, 1:5)
  lx <- list(x, x)
  do.call(merge, lx)
                      X1.5 X1.5.1
  1969-12-31 18:00:01    1      1
  1969-12-31 18:00:02    2      2
  1969-12-31 18:00:03    3      3
  1969-12-31 18:00:04    4      4
  1969-12-31 18:00:05    5      5

New behavior has column names:

  structure.1.5...Dim...c.5L..1L...index...structure.1.5..tclass...c..POSIXct...
  structure.1.5...Dim...c.5L..1L...index...structure.1.5..tclass...c..          POSIXct....1

Fixes #248. See #259.

NEWS

@@ -1,5 +1,9 @@
 Changed in xts 0.11-2:
 
+o  Make column names for merge() results with unnamed objects shorter and more
+   like zoo (#248). This also makes na.fill() much faster (#259).
+   BREAKING: This may break existing code for integer unnamed objects.
+
 o  Fix subset when 'index(x)' and 'i' contain duplicates. Thanks to Stack
    Overflow user 'scs' (https://stackoverflow.com/users/4024268/scs) for the
    report, and Philippe Verspeelt for debugging (#275).

R/merge.R

@@ -45,7 +45,19 @@ merge.xts <- function(...,
   if(is.null(suffixes)) {
     syms <- names(dots)
     syms[nchar(syms)==0] <- as.character(dots)[nchar(syms)==0]
-    if(is.null(syms)) syms <- as.character(dots)
+    if(is.null(syms)) {
+      # Based on makeNames() in merge.zoo()
+      syms <- substitute(alist(...))[-1L]
+      nm <- names(syms)
+      fixup <- if (is.null(nm)) seq_along(syms) else !nzchar(nm)
+      dep <- sapply(syms[fixup], function(x) deparse(x, nlines = 1L))
+      if(is.null(nm)) {
+        nm <- dep
+      } else if(any(fixup)) {
+        nm[fixup] <- dep
+      }
+      syms <- nm
+    }
   } else
   if(length(suffixes) != length(dots)) {
     warning("length of suffixes and does not match number of merged objects")

inst/unitTests/runit.merge.R

@@ -108,3 +108,15 @@ test.n_way_merge_on_all_types <- function() {
     checkIdentical(m, m3)
   }
 }
+
+test.shorter_colnames_for_unnamed_args <- function() {
+  X <- .xts(rnorm(10, 10), 1:10)
+
+  types <- c("double", "integer", "logical", "character", "complex")
+  for (type in types) {
+    x <- X
+    storage.mode(x) <- type
+    mx <- do.call(merge, list(x, x))
+    checkTrue(all(nchar(colnames(mx)) < 200), type)
+  }
+}