SQL query over PARTITIONED cache

classic Classic list List threaded Threaded
3 messages Options
Jeff Zemerick Jeff Zemerick
Reply | Threaded
Open this post in threaded view
|

SQL query over PARTITIONED cache

Hi, I'm new to Ignite and have a question about SQL queries over a partitioned cache. 

When I execute a SQL query with a group by and count against a LOCAL cache I get the expected results, for example:

[["test",10],["asdf",10]]

However, when I execute the same query against a PARTITIONED cache (initialized with the same data) I get some variance of:

[["test",7],["asdf",5],["asdf",5],["test",3]]

The documentation says the query is split into multiple map queries and a single reduce query. I expected the query over the PARTITIONED cache to have been reduced and match the query results from the LOCAL cache.

Am I understanding this correctly or am I missing something?

Thanks for your help.

Jeff

ilya.kasnacheev ilya.kasnacheev
Reply | Threaded
Open this post in threaded view
|

Re: SQL query over PARTITIONED cache

Hello!

This should not happen (unless you use 'collocated=true' option incorrectly).

Can you please provide your query?

Regards,
--
Ilya Kasnacheev


сб, 2 февр. 2019 г. в 19:42, Jeff Zemerick <[hidden email]>:
Hi, I'm new to Ignite and have a question about SQL queries over a partitioned cache. 

When I execute a SQL query with a group by and count against a LOCAL cache I get the expected results, for example:

[["test",10],["asdf",10]]

However, when I execute the same query against a PARTITIONED cache (initialized with the same data) I get some variance of:

[["test",7],["asdf",5],["asdf",5],["test",3]]

The documentation says the query is split into multiple map queries and a single reduce query. I expected the query over the PARTITIONED cache to have been reduced and match the query results from the LOCAL cache.

Am I understanding this correctly or am I missing something?

Thanks for your help.

Jeff

Jeff Zemerick Jeff Zemerick
Reply | Threaded
Open this post in threaded view
|

Re: SQL query over PARTITIONED cache

Ilya,

Thanks! That was the problem. 

Jeff


On Mon, Feb 4, 2019 at 1:55 AM Ilya Kasnacheev <[hidden email]> wrote:
Hello!

This should not happen (unless you use 'collocated=true' option incorrectly).

Can you please provide your query?

Regards,
--
Ilya Kasnacheev


сб, 2 февр. 2019 г. в 19:42, Jeff Zemerick <[hidden email]>:
Hi, I'm new to Ignite and have a question about SQL queries over a partitioned cache. 

When I execute a SQL query with a group by and count against a LOCAL cache I get the expected results, for example:

[["test",10],["asdf",10]]

However, when I execute the same query against a PARTITIONED cache (initialized with the same data) I get some variance of:

[["test",7],["asdf",5],["asdf",5],["test",3]]

The documentation says the query is split into multiple map queries and a single reduce query. I expected the query over the PARTITIONED cache to have been reduced and match the query results from the LOCAL cache.

Am I understanding this correctly or am I missing something?

Thanks for your help.

Jeff