fix #7364: use jvnames instead of jnames in get_shared_keys

Author: ANAMASGARD

NEWS.md

@@ -1,4 +1,8 @@
 # data.table news and updates
+## BUG FIXES
+
+1. Cartesian joins with `allow.cartesian=TRUE` after transformations using `[.data.table` would return fewer rows than expected due to incorrect `sorted` attribute being set. The issue occurred when `get_shared_keys()` used `jnames` instead of `jvnames` for index calculation, [#7364](https://github.com/Rdatatable/data.table/issues/7364). Thanks to @czeildi for reporting, @aitap for analysis, and @ANAMASGARD for fixing.
+
 
 **If you are viewing this file on CRAN, please check [latest news on GitHub](https://github.com/Rdatatable/data.table/blob/master/NEWS.md) where the formatting is also better.**
 

R/data.table.R

@@ -1449,7 +1449,7 @@ replace_dot_alias = function(e) {
           jvnames = jnames = sdvars
         else
           jnames = as.character(Filter(is.name, jsub)[-1L])
-        key_idx = chmatch(key, jnames)
+        key_idx = chmatch(key, jvnames)
         missing_keys = which(is.na(key_idx))
         if (length(missing_keys) && missing_keys[1L] == 1L) return(NULL)
         if (!length(missing_keys)) return(jvnames[key_idx])