[spark] Support 'show partitions' sql

[XuQianJin-Stars]

Purpose

fix 'show partitions' sql's bug

1. create table with partition's field type is not the same

CREATE TABLE T (a VARCHAR(10), b CHAR(10),c BIGINT,dt LONG,hh VARCHAR(4))
            PARTITIONED BY (dt, hh)
            TBLPROPERTIES ('primary-key'='a,b,dt,hh', 'bucket'='3')

2. insert some data

INSERT INTO T VALUES('a','b',1,20230816,'1132')
INSERT INTO T VALUES('a','b',1,20230816,'1133')
INSERT INTO T VALUES('a','b',1,20230816,'1134')

3. show partitions xxxx PARTITION(xxxx)

show partitions T PARTITION (hh='1132')

Tests

PaimonPartitionManagementTest

API and Format

Documentation

[YannByron]

Maybe put these in a new test class named ShowCommandTest which will contain case related to SHOW commands.

[YannByron]

try this, maybe it can work too:

import org.apache.spark.sql.execution.datasources.v2.DataSourceV2Implicits._
specOpt.asResolvedPartitionSpecs

[XuQianJin-Stars]

asResolvedPartitionSpecs

This is useful for Seq[PartitionSpec].

[YannByron]

+1 for nice catch.

[JingsongLi]

Thanks @XuQianJin-Stars , merging...