Fix locale-dependent ordering in cypher_merge tests (#2558)

The issue_1446 regression queries sort agtype labels using the database's
default collation, so their row order varies across database collations. For
example, C and en_US.utf8 order the mixed-case labels differently. Move the
incidental ordering to the enclosing SQL queries and use the C collation so
the expected output is independent of the database locale.

Author: crdv7

regress/expected/cypher_merge.out

@@ -1759,8 +1759,8 @@ $$) AS (a agtype);
 SELECT * FROM cypher('issue_1446', $$
     MATCH (n)
     RETURN labels(n) AS label, count(*) AS cnt
-    ORDER BY label
-$$) AS (label agtype, cnt agtype);
+$$) AS (label agtype, cnt agtype)
+ORDER BY label::text COLLATE "C";
  label | cnt 
 -------+-----
  ["A"] | 1
@@ -1807,8 +1807,8 @@ $$) AS (a agtype);
 SELECT * FROM cypher('issue_1446', $$
     MATCH (n)
     RETURN labels(n) AS label, count(*) AS cnt
-    ORDER BY label
-$$) AS (label agtype, cnt agtype);
+$$) AS (label agtype, cnt agtype)
+ORDER BY label::text COLLATE "C";
    label    | cnt 
 ------------+-----
  ["X"]      | 1

regress/sql/cypher_merge.sql

@@ -811,8 +811,8 @@ $$) AS (a agtype);
 SELECT * FROM cypher('issue_1446', $$
     MATCH (n)
     RETURN labels(n) AS label, count(*) AS cnt
-    ORDER BY label
-$$) AS (label agtype, cnt agtype);
+$$) AS (label agtype, cnt agtype)
+ORDER BY label::text COLLATE "C";
 SELECT * FROM cypher('issue_1446', $$
     MATCH ()-[e]->()
     RETURN count(*) AS edge_count
@@ -833,8 +833,8 @@ $$) AS (a agtype);
 SELECT * FROM cypher('issue_1446', $$
     MATCH (n)
     RETURN labels(n) AS label, count(*) AS cnt
-    ORDER BY label
-$$) AS (label agtype, cnt agtype);
+$$) AS (label agtype, cnt agtype)
+ORDER BY label::text COLLATE "C";
 SELECT * FROM cypher('issue_1446', $$
     MATCH ()-[e]->()
     RETURN count(*) AS edge_count