Fix for node reduction in case of varchar data type.

Node reduction was not working in case of varchar data types and
the reason was that the function pgxc_find_distcol_expr was not
handling the RelabelType. This patch provide the proper handling.

Also the patch changes the expected output of the test case xc_distkey
to reflect changes in explain.

2 files changed, 41 insertions, 27 deletions

diff --git a/src/backend/pgxc/plan/planner.c b/src/backend/pgxc/plan/planner.c
index e1017ce8f9..e155fa9b8f 100644
--- a/src/backend/pgxc/plan/planner.c
+++ b/src/backend/pgxc/plan/planner.c
@@ -1327,6 +1327,20 @@ pgxc_find_distcol_expr(Index varno, PartAttrNumber partAttrNum,
 
 		lexpr = linitial(op->args);
 		rexpr = lsecond(op->args);
+
+		/*
+		 * If either of the operands is a RelabelType, extract the Var in the RelabelType.
+		 * A RelabelType represents a "dummy" type coercion between two binary compatible datatypes.
+		 * If we do not handle these then our optimization does not work in case of varchar
+		 * For example if col is of type varchar and is the dist key then
+		 * select * from vc_tab where col = 'abcdefghijklmnopqrstuvwxyz';
+		 * should be shipped to one of the nodes only
+		 */
+		if (IsA(lexpr, RelabelType))
+			lexpr = ((RelabelType*)lexpr)->arg;
+		if (IsA(rexpr, RelabelType))
+			rexpr = ((RelabelType*)rexpr)->arg;
+
 		/*
 		 * If either of the operands is a Var expression, assume the other
 		 * one is distribution column expression. If none is Var check next
diff --git a/src/test/regress/expected/xc_distkey.out b/src/test/regress/expected/xc_distkey.out
index 6b6cd29628..90b0f813ae 100644
--- a/src/test/regress/expected/xc_distkey.out
+++ b/src/test/regress/expected/xc_distkey.out
@@ -39,15 +39,15 @@ select * from nu_tab where 789.412 = a;
 (1 row)
 
 explain (costs false, num_nodes true, nodes false) select * from nu_tab where a = 123.456;
-                     QUERY PLAN                      
------------------------------------------------------
- Data Node Scan (primary node count=0, node count=1)
+                                  QUERY PLAN                                   
+-------------------------------------------------------------------------------
+ Data Node Scan (primary node count=0, node count=1) on "__REMOTE_FQS_QUERY__"
 (1 row)
 
 explain (costs false, num_nodes true, nodes false) select * from nu_tab where 789.412 = a;
-                     QUERY PLAN                      
------------------------------------------------------
- Data Node Scan (primary node count=0, node count=1)
+                                  QUERY PLAN                                   
+-------------------------------------------------------------------------------
+ Data Node Scan (primary node count=0, node count=1) on "__REMOTE_FQS_QUERY__"
 (1 row)
 
 create table tx_tab(a text) distribute by modulo(a);
@@ -85,15 +85,15 @@ select * from tx_tab where 'Did the quick brown fox jump over the lazy dog?' = a
 (1 row)
 
 explain (costs false, num_nodes true, nodes false) select * from tx_tab where a = 'hello world';
-                     QUERY PLAN                      
------------------------------------------------------
- Data Node Scan (primary node count=0, node count=1)
+                                  QUERY PLAN                                   
+-------------------------------------------------------------------------------
+ Data Node Scan (primary node count=0, node count=1) on "__REMOTE_FQS_QUERY__"
 (1 row)
 
 explain (costs false, num_nodes true, nodes false) select * from tx_tab where a = 'Did the quick brown fox jump over the lazy dog?';
-                     QUERY PLAN                      
------------------------------------------------------
- Data Node Scan (primary node count=0, node count=1)
+                                  QUERY PLAN                                   
+-------------------------------------------------------------------------------
+ Data Node Scan (primary node count=0, node count=1) on "__REMOTE_FQS_QUERY__"
 (1 row)
 
 create table vc_tab(a varchar(255)) distribute by modulo(a);
@@ -128,22 +128,22 @@ select * from vc_tab where 'A quick brown fox' = a;
 (1 row)
 
 explain (costs false, num_nodes true, nodes false) select * from vc_tab where a = 'abcdefghijklmnopqrstuvwxyz';
-                     QUERY PLAN                      
------------------------------------------------------
- Data Node Scan (primary node count=0, node count=1)
+                                  QUERY PLAN                                   
+-------------------------------------------------------------------------------
+ Data Node Scan (primary node count=0, node count=1) on "__REMOTE_FQS_QUERY__"
 (1 row)
 
 explain (costs false, num_nodes true, nodes false) select * from vc_tab where a = 'A quick brown fox';
-                     QUERY PLAN                      
------------------------------------------------------
- Data Node Scan (primary node count=0, node count=1)
+                                  QUERY PLAN                                   
+-------------------------------------------------------------------------------
+ Data Node Scan (primary node count=0, node count=1) on "__REMOTE_FQS_QUERY__"
 (1 row)
 
 -- This test a bug in examine_conditions_walker where a = constant is optimized but constant = a was not
 explain (costs false, num_nodes true, nodes false) select * from vc_tab where 'A quick brown fox' = a;
-                     QUERY PLAN                      
------------------------------------------------------
- Data Node Scan (primary node count=0, node count=1)
+                                  QUERY PLAN                                   
+-------------------------------------------------------------------------------
+ Data Node Scan (primary node count=0, node count=1) on "__REMOTE_FQS_QUERY__"
 (1 row)
 
 create table f8_tab(a float8) distribute by modulo(a);
@@ -329,15 +329,15 @@ select * from i4_tab where 2147483647 = a;
 (1 row)
 
 explain (costs false, num_nodes true, nodes false) select * from i4_tab where 65530 = a;
-                     QUERY PLAN                      
------------------------------------------------------
- Data Node Scan (primary node count=0, node count=1)
+                                  QUERY PLAN                                   
+-------------------------------------------------------------------------------
+ Data Node Scan (primary node count=0, node count=1) on "__REMOTE_FQS_QUERY__"
 (1 row)
 
 explain (costs false, num_nodes true, nodes false) select * from i4_tab where a = 2147483647;
-                     QUERY PLAN                      
------------------------------------------------------
- Data Node Scan (primary node count=0, node count=1)
+                                  QUERY PLAN                                   
+-------------------------------------------------------------------------------
+ Data Node Scan (primary node count=0, node count=1) on "__REMOTE_FQS_QUERY__"
 (1 row)
 
 create table bo_tab(a bool) distribute by modulo(a);