[
https://issues.apache.org/jira/browse/IMPALA-14013?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Noémi Pap-Takács updated IMPALA-14013:
--------------------------------------
Description:
Create the table and insert values so that there will be at least 1 partition
{code:java}
> create table ice_t (i int) partitioned by (j int, s string) stored by iceberg;
> insert into ice_t select 20, 1, '2012'; {code}
Compute incremental stats for the table.
{code:java}
> compute incremental stats ice_t;
Query: compute incremental stats ice_t
+-----------------------------------------+
| summary |
+-----------------------------------------+
| Updated 1 partition(s) and 3 column(s). |
+-----------------------------------------+
Fetched 1 row(s) in 0.20s
{code}
We must specify the partition if we want to drop the incremental stats:
{code:java}
> drop incremental stats ice_t;
[Exception] ERROR: Query bc41297c26b997b2:3c2750b400000000 failed:
ParseException: Syntax error in line 1:
drop incremental stats ice_t
^
Encountered: EOF
Expected: PARTITION
CAUSED BY: Exception: Syntax error {code}
Specifying the partition (correctly) produces a NullPointerException:
{code:java}
> drop incremental stats ice_t partition (j=1, s='2012');
[Exception] ERROR: Query 2b4dcbe0b140a566:8844a7c600000000 failed:
NullPointerException: null {code}
The partitions of Iceberg tables are handled by the Iceberg library. Impala
does not have an internal representation of them, it handles Iceberg tables as
if they had only 1 HDFS partition, similarly to an unpartitioned HDFS table.
This might explain why we get an exception.
We should throw an AnalysisException with explanation if we cannot handle
incremental stats for Iceberg tables.
was:
Create the table and insert values so that there will be at least 1 partition
{code:java}
> create table ice_t (i int) partitioned by (j int, s string) stored by iceberg;
> insert into ice_t select 20, 1, '2012'; {code}
Compute incremental stats for the table.
{code:java}
> compute incremental stats ice_t;
Query: compute incremental stats ice_t
+-----------------------------------------+
| summary |
+-----------------------------------------+
| Updated 1 partition(s) and 3 column(s). |
+-----------------------------------------+
Fetched 1 row(s) in 0.20s
{code}
We must specify the partition if we want to drop the incremental stats:
{code:java}
> drop incremental stats ice_t;
[Exception] ERROR: Query bc41297c26b997b2:3c2750b400000000 failed:
ParseException: Syntax error in line 1:
drop incremental stats ice_t
^
Encountered: EOF
Expected: PARTITION
CAUSED BY: Exception: Syntax error {code}
Specifying the partition (correctly) produces a NullPointerException:
{code:java}
> drop incremental stats ice_t partition (j=1, s='2012');
[Exception] ERROR: Query 2b4dcbe0b140a566:8844a7c600000000 failed:
NullPointerException: null {code}
The partitions of Iceberg tables are handled by the Iceberg library. Impala
does not have an internal representation of them, it handles Iceberg tables as
if they had only 1 HDFS partition, similarly to an unpartitioned HDFS table.
This might explain why we get an exception.
We should throw an AnalysisException with explanation if we cannot handle
incremental stats for Iceberg tables.
> DROP INCREMENTAL STATS throws NullPointerException for Iceberg tables
> ---------------------------------------------------------------------
>
> Key: IMPALA-14013
> URL: https://issues.apache.org/jira/browse/IMPALA-14013
> Project: IMPALA
> Issue Type: Bug
> Components: Frontend
> Reporter: Noémi Pap-Takács
> Priority: Major
> Labels: impala-iceberg
>
> Create the table and insert values so that there will be at least 1 partition
>
> {code:java}
> > create table ice_t (i int) partitioned by (j int, s string) stored by
> > iceberg;
> > insert into ice_t select 20, 1, '2012'; {code}
>
> Compute incremental stats for the table.
> {code:java}
> > compute incremental stats ice_t;
> Query: compute incremental stats ice_t
> +-----------------------------------------+
> | summary |
> +-----------------------------------------+
> | Updated 1 partition(s) and 3 column(s). |
> +-----------------------------------------+
> Fetched 1 row(s) in 0.20s
> {code}
>
> We must specify the partition if we want to drop the incremental stats:
> {code:java}
> > drop incremental stats ice_t;
> [Exception] ERROR: Query bc41297c26b997b2:3c2750b400000000 failed:
> ParseException: Syntax error in line 1:
> drop incremental stats ice_t
> ^
> Encountered: EOF
> Expected: PARTITION
> CAUSED BY: Exception: Syntax error {code}
>
> Specifying the partition (correctly) produces a NullPointerException:
> {code:java}
> > drop incremental stats ice_t partition (j=1, s='2012');
> [Exception] ERROR: Query 2b4dcbe0b140a566:8844a7c600000000 failed:
> NullPointerException: null {code}
>
>
> The partitions of Iceberg tables are handled by the Iceberg library. Impala
> does not have an internal representation of them, it handles Iceberg tables
> as if they had only 1 HDFS partition, similarly to an unpartitioned HDFS
> table. This might explain why we get an exception.
> We should throw an AnalysisException with explanation if we cannot handle
> incremental stats for Iceberg tables.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
