crash on rolling join with keyed tables on=permutation of keys

[franknarf1]

I found this on SO

library(data.table)
dt <- data.table(Date = as.Date(c("2015-12-29", "2015-12-29", "2015-12-29", "2015-12-29", "2016-01-30", "2016-01    -30", "2016-01-30", "2016-01-30", "2016-02-29", "2016-02-29", "2016-02-29", "2016-02-29", "2016-03-26", "2016-03-26", "2016-03-26", "2016-03-26")), 
                 ID = c("A", "B", "C", "D", "A", "B", "C", "D", "A", "B", "C", "D", "A", "B", "C", "D"), 
                 Value = c("A201512", "B201512", "C201512", "D201512", "A201601", "B201601", "C201601", "D201601", "A201602", "B201602", "C201602", "D201602", "A201603", "B201603", "C201603", "D201603"), key = c('Date', 'ID'))
dtes <- data.table(Date=as.Date(c("2015-12-31", "2016-01-31", "2016-02-29", "2016-03-31")), key="Date")
dte <- CJ(Date=dtes$Date, ID=unique(dt$ID))
dt[dte, on=c("ID","Date"), roll=TRUE]
#
# *** caught segfault ***
#address 0xfffffffc02c53c48, cause 'memory not mapped'
#
#Traceback:
#1: bmerge(i, x, leftcols, rightcols, io, xo, roll, rollends, nomatch,     mult, ops, nqgrp, nqmaxgrp, verbose = verbose)
#2: `[.data.table`(dt, dte, on = c("ID", "Date"), roll = TRUE)
#3: dt[dte, on = c("ID", "Date"), roll = TRUE]

I'm not sure if my title covers the reason behind it...

[arunsrinivasan]

Thanks for catching and reporting this immediately. Really appreciate it!