GroupedDataFrame performs some checks on the grouped_df objects, …

…which

  prevent some issues related to corrupt `grouped_df` objects as the one
  made by rbind (#606).

NEWS.md

@@ -1,6 +1,10 @@
 # dplyr 0.2.0.9000
 
-* hashing a numeric column and an integer column was wrong (#450)
+* `GroupedDataFrame` performs some checks on the `grouped_df` objects, which 
+  prevent some issues related to corrupt `grouped_df` objects as the one 
+  made by rbind (#606).
+
+* hashing a numeric column and an integer column was wrong (#450).
 
 * `nth` now correctly promotes the result when using dates, times and factors (#509). 
 

inst/include/dplyr/GroupedDataFrame.h

@@ -36,6 +36,20 @@ namespace Rcpp {
             group_sizes = data_.attr( "group_sizes" );
             biggest_group_size  = data_.attr( "biggest_group_size" ) ;
             labels = data_.attr( "labels" );
+
+            if( !is_lazy ){
+                // check consistency of the groups
+                int rows_in_groups = sum(group_sizes) ;
+                if( data_.nrows() != rows_in_groups ){
+                    std::stringstream s ; 
+                    s << "corrupt 'grouped_df', contains "
+                      << data_.nrows()
+                      << " rows, and "
+                      << rows_in_groups
+                      << " rows in groups" ;
+                    stop(s.str()) ;
+                }
+            }
         }
 
         group_iterator group_begin() const {

tests/testthat/test-filter.r

@@ -162,3 +162,16 @@ test_that( "$ does not end call traversing. #502", {
 
 })
 
+test_that( "GroupedDataFrame checks consistency of data (#606)", {
+  df1 <- data.frame(
+   group = factor(rep(c("C", "G"), 5)),
+   value = 1:10)
+  df1 <- df1 %>% group_by(group) #df1 is now tbl
+  df2 <- data.frame(
+     group = factor(rep("G", 10)),
+     value = 11:20)
+  df3 <- rbind(df1, df2) #df2 is data.frame
+
+  expect_error( df3 %>% filter(group == "C"), "corrupt 'grouped_df', contains" )
+
+})