SQL: [tests] Fix select csv spec and enable it (#35239)

Also, replace `||` and `&&` with `OR` and `AND` as the former are not
SQL standard operators.

Author: Marios Trivyzas

x-pack/plugin/sql/qa/src/main/java/org/elasticsearch/xpack/sql/qa/jdbc/CsvSpecTestCase.java

@@ -29,6 +29,7 @@ public abstract class CsvSpecTestCase extends SpecBaseIntegrationTestCase {
     public static List<Object[]> readScriptSpec() throws Exception {
         Parser parser = specParser();
         List<Object[]> tests = new ArrayList<>();
+        tests.addAll(readScriptSpec("/select.csv-spec", parser));
         tests.addAll(readScriptSpec("/command.csv-spec", parser));
         tests.addAll(readScriptSpec("/fulltext.csv-spec", parser));
         tests.addAll(readScriptSpec("/agg.csv-spec", parser));

x-pack/plugin/sql/qa/src/main/resources/select.csv-spec

@@ -2,101 +2,110 @@
 inWithLiterals
 SELECT 1 IN (1, 2, 3), 1 IN (2, 3);
 
-  1 IN (1, 2, 3) |  1 IN (2, 3)
------------------+-------------
-true             |false
+  1 IN (1, 2, 3):b |  1 IN (2, 3):b
+-------------------+-------------
+true               |false
 ;
 
 inWithLiteralsAndFunctions
 SELECT 1 IN (2 - 1, 2, 3), abs(-1) IN (2, 3, abs(4 - 5));
 
-  1 IN (1, 2, 3) |  1 IN (2, 3)
------------------+-------------
-true             |false
+  1 IN (2 - 1, 2, 3) |  ABS(-1) IN (2, 3, ABS(4 - 5))
+---------------------+------------------------------
+true                 |true
 ;
 
 
 inWithLiteralsAndNegation
-SELECT NOT 1 IN (1, 1 + 1, 3), NOT 1 IN (2, 3);
+SELECT 1 NOT IN (1, 1 + 1, 3), 1 NOT IN (2, 3);
 
-  1 IN (1, 2, 3) |  1 IN (2, 3)
------------------+-------------
-false            |true
+  NOT(1 IN (1, 1 + 1, 3)) | NOT(1 IN (2, 3))
+--------------------------+-----------------
+false                     |true
 ;
 
-
+// Need to CAST as STRING since for boolean types the jdbc CSV translates null -> false
 inWithNullHandling
-SELECT 2 IN (1, null, 3), 3 IN (1, null, 3), null IN (1, null, 3), null IN (1, 2, 3);
+SELECT CAST(2 IN (1, null, 3) AS STRING), CAST(3 IN (1, null, 3) AS STRING), CAST(null IN (1, null, 3) AS STRING), CAST(null IN (1, 2, 3) AS STRING);
 
-  2 IN (1, null, 3) |  3 IN (1, null, 3) |  null IN (1, null, 3) |  null IN (1, 2, 3)
---------------------+--------------------+-----------------------+-------------------
-null                |true                |null                   | null
+  CAST(2 IN (1, null, 3) AS VARCHAR):s | CAST(3 IN (1, null, 3) AS VARCHAR):s |  CAST(null IN (1, null, 3) AS VARCHAR):s |  CAST(null IN (1, 2, 3) AS VARCHAR):s
+---------------------------------------+--------------------------------------+------------------------------------------+--------------------------------------
+null                                   |true                                  |null                                      |null
 ;
 
 inWithNullHandlingAndNegation
-SELECT NOT 2 IN (1, null, 3), NOT 3 IN (1, null, 3), NOT null IN (1, null, 3), NOT null IN (1, 2, 3);
+SELECT CAST(NOT 2 IN (1, null, 3) AS STRING), CAST(3 NOT IN (1, null, 3) AS STRING), CAST(NOT null IN (1, null, 3) AS STRING), CAST(null NOT IN (1, 2, 3) AS STRING);
 
-  NOT 2 IN (1, null, 3) |  NOT 3 IN (1, null, 3) |  NOT null IN (1, null, 3) |  null IN (1, 2, 3)
-------------------------+------------------------+---------------------------+--------------------
-null                    |false                   |null                       | null
+  CAST(NOT(2 IN (1, null, 3)) AS VARCHAR):s |  CAST(NOT(3 IN (1, null, 3)) AS VARCHAR):s |  CAST(NOT(null IN (1, null, 3)) AS VARCHAR):s |  CAST(NOT(null IN (1, 2, 3)) AS VARCHAR):s
+--------------------------------------------+--------------------------------------------+-----------------------------------------------+-------------------------------------------
+null                                        |false                                       |null                                           |null
 ;
 
 //
 // SELECT with IN and table columns
 //
 inWithTableColumn
-SELECT emp_no IN (10000, 10001, 10002) FROM test_emp ORDER BY 1;
-
- emp_no
--------
-10001
-10002
+SELECT emp_no IN (10000, 10001, 10002) FROM test_emp WHERE emp_no BETWEEN 10001 AND 10004 ORDER BY emp_no;
+
+ emp_no IN (10000, 10001, 10002):b
+----------------------------------
+true
+true
+false
+false
 ;
 
 inWithTableColumnAndFunction
-SELECT emp_no IN (10000, 10000 + 1, abs(-10000 - 2)) FROM test_emp;
-
- emp_no
--------
-10001
-10002
+SELECT emp_no IN (10000, 10000 + 1, abs(-10000 - 2)) FROM test_emp WHERE emp_no BETWEEN 10001 AND 10004 ORDER BY emp_no;
+
+ emp_no IN (10000, 10000 + 1, ABS(-10000 - 2)):b
+------------------------------------------------
+true
+true
+false
+false
 ;
 
 inWithTableColumnAndNegation
-SELECT emp_no NOT IN (10000, 10000 + 1, 10002) FROM test_emp ORDER BY 1 LIMIT 3;
-
- emp_no
--------
-10003
-10004
-10005
+SELECT emp_no NOT IN (10000, 10000 + 1, 10002) FROM test_emp WHERE emp_no BETWEEN 10001 AND 10004 ORDER BY emp_no;
+
+ NOT(emp_no IN (10000, 10000 + 1, 10002)):b
+-------------------------------------------
+false
+false
+true
+true
 ;
 
 inWithTableColumnAndComplexFunctions
-SELECT 1 IN (1, abs(2 - 4), 3) OR emp_no NOT IN (10000, 10000 + 1, 10002) FROM test_emp ORDER BY 1 LIMIT 3;
-
- emp_no
--------
-10003
-10004
-10005
+SELECT emp_no IN (1, abs(1 - 10002), 3) OR emp_no NOT IN (10000, 10000 + 2, 10003) FROM test_emp WHERE emp_no BETWEEN 10001 AND 10004 ORDER BY emp_no;
+
+(emp_no IN (1, ABS(1 - 10002), 3)) OR (NOT(emp_no IN (10000, 10000 + 2, 10003))):b
+----------------------------------------------------------------------------------
+true
+false
+false
+true
 ;
 
-inWithTableColumnAndNullHandling
-SELECT emp_no, birth_date in (CAST('2018-10-01T00:00:00Z' AS TIMESTAMP), CAST('1959-10-01T00:00:00Z' AS TIMESTAMP)), birth_date in (CAST('2018-10-01T00:00:00Z' AS TIMESTAMP), null, CAST('1959-10-01T00:00:00Z' AS TIMESTAMP)) FROM test_emp WHERE emp_no = 10038 OR emp_no = 10039 OR emp_no = 10040 ORDER BY 1;
 
- emp_no |  birth_date in (CAST('2018-10-01T00:00:00Z' AS TIMESTAMP), CAST('1959-10-01T00:00:00Z' AS TIMESTAMP)) |  birth_date in (CAST('2018-10-01T00:00:00Z' AS TIMESTAMP), null, CAST('1959-10-01T00:00:00Z' AS TIMESTAMP))
---------+-------------------------------------------------------------------------------------------------------+------------------------------------------------------------------------------------------------------------
-10038   | true                                                                                                  | true
-10039   | null                                                                                                  | null
-10040   | false                                                                                                 | null
+// Need to CAST as STRING since for boolean types the jdbc CSV translates null -> false
+inWithTableColumnAndNullHandling
+SELECT emp_no, CAST(languages IN (2, 3) AS STRING), CAST(languages IN (2, null, 3) AS STRING) FROM test_emp WHERE emp_no BETWEEN 10018 AND 10020 ORDER BY emp_no;
 
+ emp_no:i |  CAST(languages IN (2, 3) AS VARCHAR):s |  CAST(languages IN (2, null, 3) AS VARCHAR):s
+----------+-----------------------------------------+----------------------------------------------
+10018     |true                                     |true
+10019     |false                                    |null
+10020     |null                                     |null
+;
 
 inWithTableColumnAndNullHandlingAndNegation
-SELECT emp_no, NOT birth_date in (CAST('2018-10-01T00:00:00Z' AS TIMESTAMP), CAST('1959-10-01T00:00:00Z' AS TIMESTAMP)), NOT birth_date in (CAST('2018-10-01T00:00:00Z' AS TIMESTAMP), null, CAST('1959-10-01T00:00:00Z' AS TIMESTAMP)) FROM test_emp WHERE emp_no = 10038 OR emp_no = 10039 OR emp_no = 10040 ORDER BY 1;
+SELECT emp_no, CAST(languages NOT IN (2, 3) AS STRING), CAST(NOT languages IN (2, null, 3) AS STRING) FROM test_emp WHERE emp_no BETWEEN 10018 AND 10020 ORDER BY emp_no;
 
- emp_no |  NOT birth_date in (CAST('2018-10-01T00:00:00Z' AS TIMESTAMP), CAST('1959-10-01T00:00:00Z' AS TIMESTAMP)) |  NOT birth_date in (CAST('2018-10-01T00:00:00Z' AS TIMESTAMP), null, CAST('1959-10-01T00:00:00Z' AS TIMESTAMP))
---------+-----------------------------------------------------------------------------------------------------------+------------------------------------------------------------------------------------------------------------
-10038   | false                                                                                                     | false
-10039   | null                                                                                                      | null
-10040   | true                                                                                                      | null
+ emp_no:i |  CAST(NOT(languages IN (2, 3)) AS VARCHAR):s |  CAST(NOT(languages IN (2, null, 3)) AS VARCHAR):s
+----------+----------------------------------------------+---------------------------------------------------
+10018     |false                                         |false
+10019     |true                                          |null
+10020     |null                                          |null
+;

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/expression/predicate/logical/BinaryLogicProcessor.java

@@ -27,7 +27,7 @@ public enum BinaryLogicOperation implements PredicateBiFunction<Boolean, Boolean
                 return null;
             }
             return Boolean.logicalAnd(l.booleanValue(), r.booleanValue());
-        }, "&&"),
+        }, "AND"),
         OR((l, r) -> {
             if (Boolean.TRUE.equals(l) || Boolean.TRUE.equals(r)) {
                 return Boolean.TRUE;
@@ -36,7 +36,7 @@ public enum BinaryLogicOperation implements PredicateBiFunction<Boolean, Boolean
                 return null;
             }
             return Boolean.logicalOr(l.booleanValue(), r.booleanValue());
-        }, "||");
+        }, "OR");
 
         private final BiFunction<Boolean, Boolean, Boolean> process;
         private final String symbol;
@@ -88,4 +88,4 @@ protected void checkParameter(Object param) {
             throw new SqlIllegalArgumentException("A boolean is required; received {}", param);
         }
     }
-}
+}

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/expression/predicate/operator/comparison/In.java

@@ -90,7 +90,7 @@ public Boolean fold() {
 
     @Override
     public String name() {
-        StringJoiner sj = new StringJoiner(", ", " IN(", ")");
+        StringJoiner sj = new StringJoiner(", ", " IN (", ")");
         list.forEach(e -> sj.add(Expressions.name(e)));
         return Expressions.name(value) + sj.toString();
     }

x-pack/plugin/sql/src/test/java/org/elasticsearch/xpack/sql/planner/QueryTranslatorTests.java

@@ -197,7 +197,7 @@ public void testTranslateInExpressionInvalidValues_WhereClause() {
         assertFalse(condition.foldable());
         SqlIllegalArgumentException ex = expectThrows(SqlIllegalArgumentException.class, () -> QueryTranslator.toQuery(condition, false));
         assertEquals("Line 1:52: Comparisons against variables are not (currently) supported; " +
-                "offender [keyword] in [keyword IN(foo, bar, keyword)]", ex.getMessage());
+                "offender [keyword] in [keyword IN (foo, bar, keyword)]", ex.getMessage());
     }
 
     public void testTranslateInExpression_WhereClause_Painless() {