:= suppressed in a more reliable way not depending on trigger depth. …

…Both if(TRUE)DT[,:=] and `:=` in knitr should now be suppressed. See README for one downside, though. Closes #869.
Rdatatable · Oct 8, 2014 · 342d6ba · 342d6ba
1 parent e011351
commit 342d6ba
Show file tree

Hide file tree

Showing 4 changed files with 75 additions and 31 deletions.
diff --git a/R/data.table.R b/R/data.table.R
@@ -57,13 +57,7 @@ dim.data.table <- function(x) {
 
 .global = new.env()  # thanks to: http://stackoverflow.com/a/12605694/403310
 setPackageName("data.table",.global)
-.global$print = TRUE
-.global$depthtrigger =
-    if (exists(".global",.GlobalEnv)) {
-        9L # this is development
-    } else {
-        3L # normal value when package is loaded in 2.14+ (where functions are compiled in namespace). data.table depends on R >= 2.14.0
-    }
+.global$print = ""
 
 .SD = .N = .I = .GRP = .BY = NULL
 # These are exported to prevent NOTEs from R CMD check, and checkUsage via compiler.
@@ -78,11 +72,15 @@ print.data.table = function(x,
     nrows=getOption("datatable.print.nrows"), # (100) under this the whole (small) table is printed, unless topn is provided
     row.names = TRUE, ...)
 {
-    if (!.global$print) {
-        #  := in [.data.table sets print=FALSE, when appropriate, to suppress := autoprinting at the console
-        .global$print = TRUE
+    if (.global$print != "" &&
+        length(SYS<-last(sys.calls()))>=2 &&
+        typeof(SYS[[2]]) == "list" &&   # auto-printing; e.g. not explicit print.  When explicit print, this is symbol or language
+        address(x) == .global$print ) {
+        #  := in [.data.table sets print=address(x), when appropriate, to suppress next autoprint at the console. See FAQ 2.22
+        .global$print = ""
         return(invisible())
     }
+    .global$print = ""
     if (!is.numeric(nrows)) nrows = 100L
     if (!is.infinite(nrows)) nrows = as.integer(nrows)
     if (nrows <= 0L) return(invisible())   # ability to turn off printing
@@ -357,10 +355,10 @@ chmatch2 <- function(x, table, nomatch=NA_integer_) {
     if (!is.logical(which) || length(which)>1) stop("'which' must be a logical vector length 1. Either FALSE, TRUE or NA.")
     if ((isTRUE(which)||is.na(which)) && !missing(j)) stop("'which' is ",which," (meaning return row numbers) but 'j' is also supplied. Either you need row numbers or the result of j, but only one type of result can be returned.")
     if (!is.na(nomatch) && is.na(which)) stop("which=NA with nomatch=0 would always return an empty vector. Please change or remove either which or nomatch.")
+    .global$print=""
     if (missing(i) && missing(j)) {
         # ...[] == oops at console, forgot print(...)
         # or some kind of dynamic construction that has edge case of no contents inside [...]
-        .global$print=TRUE
         return(x)
     }
     if (!with && missing(j)) stop("j must be provided when with=FALSE")
@@ -878,11 +876,10 @@ chmatch2 <- function(x, table, nomatch=NA_integer_) {
             suppPrint = identity
             if (length(av) && av[1L] == ":=") {
                 if (identical(attr(x,".data.table.locked"),TRUE)) stop(".SD is locked. Using := in .SD's j is reserved for possible future use; a tortuously flexible way to modify by group. Use := in j directly to modify by group by reference.")
-                if (Cstack_info()[["eval_depth"]] <= .global$depthtrigger) {
-                    suppPrint = function(x) { .global$print=FALSE; x }
-                    # Suppress print when returns ok not on error, bug #2376. Thanks to: http://stackoverflow.com/a/13606880/403310
-                    # All appropriate returns following this point are wrapped i.e. return(suppPrint(x)).
-                }
+                suppPrint = function(x) { .global$print=address(x); x }
+                # Suppress print when returns ok not on error, bug #2376. Thanks to: http://stackoverflow.com/a/13606880/403310
+                # All appropriate returns following this point are wrapped; i.e. return(suppPrint(x)).
+
                 # FR #4996 - verbose message and return when a join matches nothing with `:=` in j
                 if (byjoin & !notjoin) {
                     # Note: !notjoin is here only until the notjoin is implemented as a "proper" byjoin

diff --git a/README.md b/README.md
@@ -11,11 +11,13 @@
 
 #### NEW FEATURES
 
+  1. `if (TRUE) DT[,LHS:=RHS]` now doesn't print (thanks to Jureiss, [#869](https://github.com/Rdatatable/data.table/issues/869) and `:=` should not longer print in knitr. To get this to work we've had to live with one downside: if a `:=` is used inside a function with no `DT[]` before the end of the function, then the next time `DT` is typed at the prompt, nothing will be printed. A repeated `DT` will print. To avoid this: include a `DT[]` after the last `:=` in your function. If that is not possible (e.g., it's not a function you can change) then `print(DT)` and `DT[]` at the prompt are guaranteed to print. As before, adding an extra `[]` on the end of `:=` query is a recommended idiom to update and then print; e.g. `> DT[,foo:=3L][]`
+
 #### BUG FIXES
 
   1. `as.data.table.list` with list input having 0-length items, e.g. `x = list(a=integer(0), b=3:4)`. `as.data.table(x)` recycles item `a` with `NA`s to fit the length of the longer column `b` (length=2), as before now, but with an additional warning message that the item has been recycled with `NA`. Closes [#847](https://github.com/Rdatatable/data.table/issues/847). Thanks to @tvinodr for the report. This was a regression from 1.9.2.
 
-  2. In `DT[i, j]` when `i` returns all `FALSE`, some edge cases where `j` contained length-0 values (ex: `integer(0)`) did not return an empty data.table, as it should. Closes [#758](https://github.com/Rdatatable/data.table/issues/758) and [#813](https://github.com/Rdatatable/data.table/issues/813). Thanks to @tunaaa and @nigmastar for the nice reproducible reports. 
+  2. `DT[i, j]` when `i` returns all `FALSE` and `j` contains some length-0 values (ex: `integer(0)`) now returns an empty data.table as it should. Closes [#758](https://github.com/Rdatatable/data.table/issues/758) and [#813](https://github.com/Rdatatable/data.table/issues/813). Thanks to @tunaaa and @nigmastar for the nice reproducible reports. 
 
   3. `allow.cartesian` is ignored during joins when:  
       * `i` has no duplicates and `mult="all"`. Closes [#742](https://github.com/Rdatatable/data.table/issues/742). Thanks to @nigmastar for the report.  

diff --git a/tests/autoprint.R b/tests/autoprint.R
@@ -8,22 +8,35 @@ DT[2,a:=3L]                           # no
 DT                                    # yes
 print(DT[2,a:=4L])                    # yes
 print(DT)                             # yes
-(function(){DT[2,a:=5L];NULL})()      # no
-DT                                    # yes
+if (TRUE) DT[2,a:=5L]                 # no. used to print before v1.9.5
+if (TRUE) if (TRUE) DT[2,a:=6L]       # no. used to print before v1.9.5
+(function(){DT[2,a:=5L];NULL})()      # print NULL
+DT                                    # no (from v1.9.5+). := suppresses next auto print (can't distinguish just "DT" symbol alone at the prompt)
+DT                                    # yes. 2nd time needed, or solutions below
+(function(){DT[2,a:=5L];NULL})()      # print NULL
+DT[]                                  # yes. guaranteed print
+(function(){DT[2,a:=5L];NULL})()      # print NULL
+print(DT)                             # yes. guaranteed print
+(function(){DT[2,a:=5L][];NULL})()    # print NULL
+DT                                    # yes. i) function needs to add [] after last one, so that "DT" alone is guaranteed anyway
+(function(){DT[2,a:=5L];DT[];NULL})() # print NULL
+DT                                    # yes. ii) or as a separate DT[] after the last := inside the function
+DT2 = data.table(b=3:4)               # no
+(function(){DT[2,a:=6L];DT2[1,b:=7L];NULL})()
+DT                                    # yes. last := was on DT2 not DT
 {DT[2,a:=6L];invisible()}             # no
 print(DT)                             # yes
 (function(){print(DT[2,a:=7L]);print(DT);invisible()})()    # yes*2
 {print(DT[2,a:=8L]);print(DT);invisible()}                  # yes*2
 DT[1][,a:=9L]      # no (was too tricky to detect that DT[1] is a new object). Simple rule is that := always doesn't print
-DT[2,a:=10L][1]    # yes (because eval depth is above trigger in the := here via nested `[.data.table` calls, iiuc).
+DT[2,a:=10L][1]                       # yes
 DT[1,a:=10L][1,a:=10L]                # no
 DT[,a:=as.integer(a)]                 # no
 DT[1,a:=as.integer(a)]                # no
 DT[1,a:=10L][]                        # yes. ...[] == oops, forgot print(...)
 
 # Test that error in := doesn't suppress next valid print, bug #2376
-assign("depthtrigger", 20L, data.table:::.global)   # try() adds 10 levels of depth. But we need try() otherwise this script would stop.
 try(DT[,foo:=ColumnNameTypo])         # error: not found.
-DT                                    # yes  (if we didn't change depthtrigger above this would print for depth reason and not test it)
+DT                                    # yes
 DT                                    # yes
 
diff --git a/tests/autoprint.Rout.save b/tests/autoprint.Rout.save
@@ -1,7 +1,6 @@
 
-R version 2.15.1 (2012-06-22) -- "Roasted Marshmallows"
-Copyright (C) 2012 The R Foundation for Statistical Computing
-ISBN 3-900051-07-0
+R version 3.1.1 (2014-07-10) -- "Sock it to Me"
+Copyright (C) 2014 The R Foundation for Statistical Computing
 Platform: x86_64-pc-linux-gnu (64-bit)
 
 R is free software and comes with ABSOLUTELY NO WARRANTY.
@@ -41,12 +40,46 @@ Loading required package: data.table
    a
 1: 1
 2: 4
-> (function(){DT[2,a:=5L];NULL})()      # no
+> if (TRUE) DT[2,a:=5L]                 # no. used to print before v1.9.5
+> if (TRUE) if (TRUE) DT[2,a:=6L]       # no. used to print before v1.9.5
+> (function(){DT[2,a:=5L];NULL})()      # print NULL
 NULL
-> DT                                    # yes
+> DT                                    # no (from v1.9.5+). := suppresses next auto print (can't distinguish just "DT" symbol alone at the prompt)
+> DT                                    # yes. 2nd time needed, or solutions below
+   a
+1: 1
+2: 5
+> (function(){DT[2,a:=5L];NULL})()      # print NULL
+NULL
+> DT[]                                  # yes. guaranteed print
+   a
+1: 1
+2: 5
+> (function(){DT[2,a:=5L];NULL})()      # print NULL
+NULL
+> print(DT)                             # yes. guaranteed print
+   a
+1: 1
+2: 5
+> (function(){DT[2,a:=5L][];NULL})()    # print NULL
+NULL
+> DT                                    # yes. i) function needs to add [] after last one, so that "DT" alone is guaranteed anyway
+   a
+1: 1
+2: 5
+> (function(){DT[2,a:=5L];DT[];NULL})() # print NULL
+NULL
+> DT                                    # yes. ii) or as a separate DT[] after the last := inside the function
    a
 1: 1
 2: 5
+> DT2 = data.table(b=3:4)               # no
+> (function(){DT[2,a:=6L];DT2[1,b:=7L];NULL})()
+NULL
+> DT                                    # yes. last := was on DT2 not DT
+   a
+1: 1
+2: 6
 > {DT[2,a:=6L];invisible()}             # no
 > print(DT)                             # yes
    a
@@ -67,7 +100,7 @@ NULL
 1: 1
 2: 8
 > DT[1][,a:=9L]      # no (was too tricky to detect that DT[1] is a new object). Simple rule is that := always doesn't print
-> DT[2,a:=10L][1]    # yes (because eval depth is above trigger in the := here via nested `[.data.table` calls, iiuc).
+> DT[2,a:=10L][1]                       # yes
    a
 1: 1
 > DT[1,a:=10L][1,a:=10L]                # no
@@ -79,10 +112,9 @@ NULL
 2: 10
 > 
 > # Test that error in := doesn't suppress next valid print, bug #2376
-> assign("depthtrigger", 20L, data.table:::.global)   # try() adds 10 levels of depth. But we need try() otherwise this script would stop.
 > try(DT[,foo:=ColumnNameTypo])         # error: not found.
 Error in eval(expr, envir, enclos) : object 'ColumnNameTypo' not found
-> DT                                    # yes  (if we didn't change depthtrigger above this would print for depth reason and not test it)
+> DT                                    # yes
     a
 1: 10
 2: 10
@@ -94,4 +126,4 @@ Error in eval(expr, envir, enclos) : object 'ColumnNameTypo' not found
 > 
 > proc.time()
    user  system elapsed 
-  1.740   0.096   1.810 
+   3.14    0.10    3.22