Frank Yin created SPARK-32059:
---------------------------------
Summary: Nested Schema Pruning not Working in Window Functions
Key: SPARK-32059
URL: https://issues.apache.org/jira/browse/SPARK-32059
Project: Spark
Issue Type: Bug
Components: SQL
Affects Versions: 3.0.0
Reporter: Frank Yin
Using tables and data structures in `SchemaPruningSuite.scala`
```
case class FullName(first: String, middle: String, last: String)
case class Company(name: String, address: String)
case class Employer(id: Int, company: Company)
case class Contact(
id: Int,
name: FullName,
address: String,
pets: Int,
friends: Array[FullName] = Array.empty,
relatives: Map[String, FullName] = Map.empty,
employer: Employer = null,
relations: Map[FullName, String] = Map.empty)
case class Department(
depId: Int,
depName: String,
contactId: Int,
employer: Employer)
```
The query to run: `
select a.name.first from (select row_number() over (partition by address order
by id desc) as __rank, contacts.* from contacts) a where a.name.first = 'A' AND
a.__rank = 1`
The current physical plan:
```
== Physical Plan ==
*(3) Project [name#46.first AS first#74]
+- *(3) Filter (((isnotnull(name#46) AND isnotnull(__rank#71)) AND
(name#46.first = A)) AND (__rank#71 = 1))
+- Window [row_number() windowspecdefinition(address#47, id#45 DESC NULLS
LAST, specifiedwindowframe(RowFrame, unboundedpreceding$(), currentrow$())) AS
__rank#71], [address#47], [id#45 DESC NULLS LAST]
+- *(2) Sort [address#47 ASC NULLS FIRST, id#45 DESC NULLS LAST], false, 0
+- Exchange hashpartitioning(address#47, 5), true, [id=#52]
+- *(1) Project [id#45, name#46, address#47]
+- FileScan parquet [id#45,name#46,address#47,p#53] Batched: false,
DataFilters: [], Format: Parquet, Location:
InMemoryFileIndex[file:/private/var/folders/_c/4r2j33dd14n9ldfc2xqyzs400000gn/T/spark-85d173af-42...,
PartitionFilters: [], PushedFilters: [], ReadSchema:
struct<id:int,name:struct<first:string,middle:string,last:string>,address:string>
```
The desired physical plan:
```
== Physical Plan ==
*(3) Project [_gen_alias_77#77 AS first#74]
+- *(3) Filter (((isnotnull(_gen_alias_77#77) AND isnotnull(__rank#71)) AND
(_gen_alias_77#77 = A)) AND (__rank#71 = 1))
+- Window [row_number() windowspecdefinition(address#47, id#45 DESC NULLS
LAST, specifiedwindowframe(RowFrame, unboundedpreceding$(), currentrow$())) AS
__rank#71], [address#47], [id#45 DESC NULLS LAST]
+- *(2) Sort [address#47 ASC NULLS FIRST, id#45 DESC NULLS LAST], false, 0
+- Exchange hashpartitioning(address#47, 5), true, [id=#52]
+- *(1) Project [id#45, name#46.first AS _gen_alias_77#77, address#47]
+- FileScan parquet [id#45,name#46,address#47,p#53] Batched: false,
DataFilters: [], Format: Parquet, Location:
InMemoryFileIndex[file:/private/var/folders/_c/4r2j33dd14n9ldfc2xqyzs400000gn/T/spark-c64e0b29-d9...,
PartitionFilters: [], PushedFilters: [], ReadSchema:
struct<id:int,name:struct<first:string>,address:string>
```
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
