[
https://issues.apache.org/jira/browse/CARBONDATA-2544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16524445#comment-16524445
]
xubo245 commented on CARBONDATA-2544:
-------------------------------------
It's work fine in cluster：
{code:java}
0: jdbc:hive2://hadoop1:10000> select country,sum(salary) from test20 group by
country;
+----------+--------------+--+
| country | sum(salary) |
+----------+--------------+--+
| USA | 23 |
+----------+--------------+--+
1 row selected (1.226 seconds)
0: jdbc:hive2://hadoop1:10000> select country,sum(salary) from test20 where
country='USA' group by country;
+----------+--------------+--+
| country | sum(salary) |
+----------+--------------+--+
| USA | 23 |
+----------+--------------+--+
1 row selected (1.655 seconds)
0: jdbc:hive2://hadoop1:10000> select country,sum(salary) from test20 where
country='USA' group by country;
+----------+--------------+--+
| country | sum(salary) |
+----------+--------------+--+
| USA | 23 |
+----------+--------------+--+
1 row selected (0.92 seconds)
0: jdbc:hive2://hadoop1:10000> explain select country,sum(salary) from test20
where country='USA' group by country;
+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+--+
|
plan
|
+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+--+
| == CarbonData Profiler ==
Table Scan on datamv20_table
- total blocklets: 1
- filter: (test20_country <> null and test20_country = USA)
- pruned by Main DataMap
- skipped blocklets: 0
|
| == Physical Plan ==
*HashAggregate(keys=[country#223], functions=[sum(sum(salary)#224L)])
+- Exchange hashpartitioning(country#223, 200)
+- *HashAggregate(keys=[country#223],
functions=[partial_sum(sum(salary)#224L)])
+- *HashAggregate(keys=[test20_country#103],
functions=[sum(sum_salary#104L)])
+- Exchange hashpartitioning(test20_country#103, 200)
+- *HashAggregate(keys=[test20_country#103],
functions=[partial_sum(sum_salary#104L)])
+- *BatchedScan CarbonDatasourceHadoopRelation [ Database name
:default, Table name :datamv20_table, Schema
:Some(StructType(StructField(test20_country,StringType,true),
StructField(sum_salary,LongType,true))) ]
default.datamv20_table[test20_country#103,sum_salary#104L] PushedFilters:
[IsNotNull(test20_country), EqualTo(test20_country,USA)] |
+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+--+
2 rows selected (0.318 seconds)
{code}
> [MV] Wrong data displayed with Filter
> --------------------------------------
>
> Key: CARBONDATA-2544
> URL: https://issues.apache.org/jira/browse/CARBONDATA-2544
> Project: CarbonData
> Issue Type: Bug
> Reporter: Babulal
> Assignee: xubo245
> Priority: Major
>
> spark.sql("drop table if exists test1")
> spark.sql("create table test1( name string,country string,age int,salary
> int) stored by 'carbondata' ")
> spark.sql("insert into test1 select 'name1','USA',12,23")
> spark.sql("create datamap datamv2 using 'mv' as select country,sum(salary)
> from test1 group by country").show()
> spark.sql("rebuild datamap datamv2")
> spark.sql("select country,sum(salary) from test1 group by
> country").show(200,false)
> +--------+----------+
> |country|sum(salary)|
> +--------+----------+
> |USA|23|
> +--------+----------+
>
> spark.sql("select country,sum(salary) from test1 where country='USA' group by
> country").show(200,false)
> +--------+----------+
> |country|sum(salary)|
> +--------+----------+
> +--------+----------+
>
> This is because, select query formation is wrong , filter value is changed to
> lowercase
> 2018-05-27 00:20:16 INFO CarbonSparkSqlParser:54 - Parsing command: select
> preAGG() as preAgg, gen_subsumer_0.`country`, gen_subsumer_0.`sum(salary)` as
> `sum(salary)`
> from
> (select test1.`country`, sum(cast(test1.`salary` as bigint)) as
> `sum(salary)`
> from
> test1
> group by test1.`country`) gen_subsumer_0
> where
> (gen_subsumer_0.`country` = 'usa')
>
>
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
