Github user yhuai commented on a diff in the pull request:

    https://github.com/apache/spark/pull/9399#discussion_r43719436
  
    --- Diff: 
sql/core/src/test/scala/org/apache/spark/sql/sources/FilteredScanSuite.scala ---
    @@ -202,51 +232,60 @@ class FilteredScanSuite extends DataSourceTest with 
SharedSQLContext {
         "SELECT a, b, c FROM oneToTenFiltered WHERE c like '%eE%'",
         Seq(Row(5, 5 * 2, "e" * 5 + "E" * 5)))
     
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE A = 1", 1)
    -  testPushDown("SELECT a FROM oneToTenFiltered WHERE A = 1", 1)
    -  testPushDown("SELECT b FROM oneToTenFiltered WHERE A = 1", 1)
    -  testPushDown("SELECT a, b FROM oneToTenFiltered WHERE A = 1", 1)
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE a = 1", 1)
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE 1 = a", 1)
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE A = 1", 1, Set("a", 
"b", "c"))
    +  testPushDown("SELECT a FROM oneToTenFiltered WHERE A = 1", 1, Set("a"))
    +  testPushDown("SELECT b FROM oneToTenFiltered WHERE A = 1", 1, Set("b"))
    +  testPushDown("SELECT a, b FROM oneToTenFiltered WHERE A = 1", 1, 
Set("a", "b"))
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE a = 1", 1, Set("a", 
"b", "c"))
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE 1 = a", 1, Set("a", 
"b", "c"))
    +
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE a > 1", 9, Set("a", 
"b", "c"))
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE a >= 2", 9, Set("a", 
"b", "c"))
     
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE a > 1", 9)
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE a >= 2", 9)
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE 1 < a", 9, Set("a", 
"b", "c"))
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE 2 <= a", 9, Set("a", 
"b", "c"))
     
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE 1 < a", 9)
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE 2 <= a", 9)
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE 1 > a", 0, Set("a", 
"b", "c"))
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE 2 >= a", 2, Set("a", 
"b", "c"))
     
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE 1 > a", 0)
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE 2 >= a", 2)
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE a < 1", 0, Set("a", 
"b", "c"))
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE a <= 2", 2, Set("a", 
"b", "c"))
     
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE a < 1", 0)
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE a <= 2", 2)
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE a > 1 AND a < 10", 8, 
Set("a", "b", "c"))
     
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE a > 1 AND a < 10", 8)
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE a IN (1,3,5)", 3, 
Set("a", "b", "c"))
     
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE a IN (1,3,5)", 3)
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE a = 20", 0, Set("a", 
"b", "c"))
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE b = 1", 10, Set("a", 
"b", "c"))
     
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE a = 20", 0)
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE b = 1", 10)
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE a < 5 AND a > 1", 3, 
Set("a", "b", "c"))
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE a < 3 OR a > 8", 4, 
Set("a", "b", "c"))
    +  testPushDown("SELECT * FROM oneToTenFiltered WHERE NOT (a < 6)", 5, 
Set("a", "b", "c"))
     
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE a < 5 AND a > 1", 3)
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE a < 3 OR a > 8", 4)
    -  testPushDown("SELECT * FROM oneToTenFiltered WHERE NOT (a < 6)", 5)
    +  testPushDown("SELECT a, b, c FROM oneToTenFiltered WHERE c like 'c%'", 
1, Set("a", "b", "c"))
    +  testPushDown("SELECT a, b, c FROM oneToTenFiltered WHERE c like 'C%'", 
0, Set("a", "b", "c"))
     
    -  testPushDown("SELECT a, b, c FROM oneToTenFiltered WHERE c like 'c%'", 1)
    -  testPushDown("SELECT a, b, c FROM oneToTenFiltered WHERE c like 'C%'", 0)
    +  testPushDown("SELECT a, b, c FROM oneToTenFiltered WHERE c like '%D'", 
1, Set("a", "b", "c"))
    +  testPushDown("SELECT a, b, c FROM oneToTenFiltered WHERE c like '%d'", 
0, Set("a", "b", "c"))
     
    -  testPushDown("SELECT a, b, c FROM oneToTenFiltered WHERE c like '%D'", 1)
    -  testPushDown("SELECT a, b, c FROM oneToTenFiltered WHERE c like '%d'", 0)
    +  testPushDown("SELECT a, b, c FROM oneToTenFiltered WHERE c like '%eE%'", 
1, Set("a", "b", "c"))
    +  testPushDown("SELECT a, b, c FROM oneToTenFiltered WHERE c like '%Ee%'", 
0, Set("a", "b", "c"))
     
    -  testPushDown("SELECT a, b, c FROM oneToTenFiltered WHERE c like '%eE%'", 
1)
    -  testPushDown("SELECT a, b, c FROM oneToTenFiltered WHERE c like '%Ee%'", 
0)
    +  testPushDown("SELECT c FROM oneToTenFiltered WHERE c = 'aaaaaAAAAA'", 1, 
Set("c"))
    +  testPushDown("SELECT c FROM oneToTenFiltered WHERE c IN ('aaaaaAAAAA', 
'foo')", 1, Set("c"))
     
    -  testPushDown("SELECT c FROM oneToTenFiltered WHERE c = 'aaaaaAAAAA'", 1)
    -  testPushDown("SELECT c FROM oneToTenFiltered WHERE c IN ('aaaaaAAAAA', 
'foo')", 1)
    +  // Columns only referenced by UDF filter must be required, as UDF 
filters can't be pushed down.
    +  testPushDown("SELECT c FROM oneToTenFiltered WHERE udf_even(A)", 10, 
Set("a", "c"))
     
    -  def testPushDown(sqlString: String, expectedCount: Int): Unit = {
    +  def testPushDown(
    --- End diff --
    
    Let's also add checks to make sure the `Filter` operator added by Spark SQL 
only contains unhandled predicates, unconvertible predicates, and predicates 
involving partition columns.


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to