[
https://issues.apache.org/jira/browse/IGNITE-18223?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vladimir Steshin updated IGNITE-18223:
--------------------------------------
Description:
To replicate data through CDC, SQL/CDC documentation should mention that both
SQL tables on the source and destination nodes have to have the same
'value_type' like:
{code:sql}
CREATE TABLE Person (id int primary key, varchar name) with
cache_name='Person', value_type='SQL_PERSON_TABLE_TYPE'
{code}
Otherwise, successfully replicated data is not seen by SQL queries on the
receiver node. This happens because the recipient node has its own generated
binary type name for its related table. Replication of schema or any settings
aren't supported yet.
Binary type name is randomly generated for SQL table because we do not delete
binary types at all. Type name based only on table name/schema can breake:
{code:sql}
sql("create table tst(id int)");
sql("insert into tst values (1)");
sql("drop table tst");
sql("create table tst(id varchar)");
sql("insert into tst values ('1')");
{code}
was:
To replicate data through CDC, SQL/CDC documentation should mention that both
SQL tables on the source and destination nodes have to have the same
'value_type' like:
{code:sql}
CREATE TABLE Person (id int primary key, varchar name) with
cache_name='Person', value_type='SQL_PERSON_TABLE_TYPE'
{code}
Otherwise, successfully replicated data is not seen by SQL queries on the
receiver node. This happens because the recipient node has its own generated
binary type name for its related table. Replication of schema or any settings
aren't supported yet.
Binary type name is randomly generated for SQL table because we do not delete
binary types at all. Type name based only on table name/schema can brake:
{code:sql}
sql("create table tst(id int)");
sql("insert into tst values (1)");
sql("drop table tst");
sql("create table tst(id varchar)");
sql("insert into tst values ('1')");
{code}
> Document same SQL table value_type for CDC.
> -------------------------------------------
>
> Key: IGNITE-18223
> URL: https://issues.apache.org/jira/browse/IGNITE-18223
> Project: Ignite
> Issue Type: Improvement
> Reporter: Vladimir Steshin
> Priority: Major
> Labels: ise
>
> To replicate data through CDC, SQL/CDC documentation should mention that both
> SQL tables on the source and destination nodes have to have the same
> 'value_type' like:
> {code:sql}
> CREATE TABLE Person (id int primary key, varchar name) with
> cache_name='Person', value_type='SQL_PERSON_TABLE_TYPE'
> {code}
> Otherwise, successfully replicated data is not seen by SQL queries on the
> receiver node. This happens because the recipient node has its own generated
> binary type name for its related table. Replication of schema or any settings
> aren't supported yet.
> Binary type name is randomly generated for SQL table because we do not delete
> binary types at all. Type name based only on table name/schema can breake:
> {code:sql}
> sql("create table tst(id int)");
> sql("insert into tst values (1)");
> sql("drop table tst");
> sql("create table tst(id varchar)");
> sql("insert into tst values ('1')");
> {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
