Closes #3106, better error message for unmatched patterns() in melt()

NEWS.md

@@ -15,6 +15,8 @@
 
 2. Column names that look like expressions (e.g. `"a<=colB"`) caused an error when used in `on=` even when wrapped with backticks, [#3092](https://github.com/Rdatatable/data.table/issues/3092). Additionally, `on=` now supports white spaces around operators; e.g. `on = "colA == colB"`. Thanks to @mt1022 for reporting and to @MarkusBonsch for fixing.
 
+3. Unmatched `patterns` in `measure.vars` fail early and with feedback, [#3106](https://github.com/Rdatatable/data.table/issues/3092).
+
 #### NOTES
 
 1. When data.table first loads it now checks the DLL's MD5. This is to detect installation issues on Windows when you upgrade and i) the DLL is in use by another R session and ii) the CRAN source version > CRAN binary binary which happens just after a new release (R prompts users to install from source until the CRAN binary is available). This situation can lead to a state where the package's new R code calls old C code in the old DLL; [R#17478](https://bugs.r-project.org/bugzilla/show_bug.cgi?id=17478), [#3056](https://github.com/Rdatatable/data.table/issues/3056). This broken state can persist until, hopefully, you experience a strange error caused by the mismatch. Otherwise, wrong results may occur silently. This situation applies to any R package with compiled code not just data.table, is Windows-only, and is long-standing. It has only recently been understood as it typically only occurs during the few days after each new release until binaries are available on CRAN. Thanks to Gabor Csardi for the suggestion to use `tools::checkMD5sums()`.

R/fmelt.R

@@ -42,6 +42,10 @@ melt.data.table <- function(data, id.vars, measure.vars, variable.name = "variab
     } else cols = names(data)
     pats = lapply(measure.sub, eval, parent.frame())
     measure.vars = patterns(pats, cols=cols)
+    # replace with lengths when R 3.2.0 dependency arrives
+    if (length(idx <- which(sapply(measure.vars, length) == 0L)))
+      stop('Pattern', if (length(idx) > 1L) 's', ' not found: [',
+           paste(pats[idx], collapse = ', '), ']')
   }
   if (is.list(measure.vars) && length(measure.vars) > 1L) {
     meas.nm = names(measure.vars)

inst/tests/tests.Rraw

@@ -12328,6 +12328,13 @@ test(1951.4, d1[d2, nomatch=3], error="nomatch= must be either NA or NULL .or 0
 test(1952.1, d1[a==2, which=3], error="which= must be a logical vector length 1. Either FALSE, TRUE or NA.")
 test(1952.2, d1[a==2, 2, which=TRUE], error="which==TRUE.*but j is also supplied")
 
+# 3106 -- melt patterns don't match any columns
+DT = data.table(id = 1:3, a1 = rnorm(3), a2 = rnorm(3))
+test(1953.1, melt(DT, id.vars = 'id', measure.vars = patterns(a = 'a', b = 'b')),
+     error = 'Pattern not found')
+test(1953.2, melt(DT, id.vars = 'id',
+                  measure.vars = patterns(a = 'a', b = 'b', c = 'c')),
+     error = 'Patterns not found')
 
 ###################################
 #  Add new tests above this line  #