nizhikov commented on code in PR #13685:
URL: https://github.com/apache/ignite/pull/13685#discussion_r4423587101


##########
docs/_docs/SQL/sql-calcite.adoc:
##########
@@ -1294,3 +1309,76 @@ If neither option is specified, Ignite uses the 
transaction timeout.
 `SELECT ... FOR UPDATE` is supported only for plain `SELECT` queries over 
cache-based tables and joins of such tables.
 It is not supported for `DISTINCT`, `GROUP BY`, `HAVING`, aggregate-only 
result sets, system views, or table functions.
 ====
+
+== Legacy H2 Engine [[legacy-h2-engine]]
+
+The SQL engine based on the H2 database is provided by the `ignite-indexing` 
module. The module is shipped in the `libs` folder of the binary distribution 
next to `ignite-calcite`, so a node started with the `ignite.{sh|bat}` script 
has both engines: the Calcite-based engine is the default one and the H2 engine 
is available by name, see <<Configuring Query Engines>>. If you start Ignite 
from your application and manage its dependencies with Maven, add the 
`ignite-indexing` dependency to get the H2 engine.
+
+To make the H2 engine the default engine again, configure it explicitly with 
`IndexingQueryEngineConfiguration.setDefault(true)`, as shown in <<Configuring 
Query Engines>>. This is the simplest way to keep the previous behaviour after 
an upgrade.
+
+[WARNING]
+====
+Every node of the cluster, including thick client nodes and the nodes embedded 
into an application, must have the same set of engine modules as the server 
nodes.
+====
+
+=== Features Implemented by the H2 Engine Only
+
+The following features work only when the query is executed by the H2 engine: 
the `ignite-indexing` module must be on the classpath of every node, and the H2 
engine must be selected for the query or the connection, or configured as the 
default engine.
+
+* The link:sql-reference/operational-commands#copy[COPY] and 
link:sql-reference/operational-commands#set-streaming[SET STREAMING] commands 
and the streaming modes of the JDBC and ODBC drivers.
+* link:SQL/JDBC/jdbc-driver#partition-awareness[Partition awareness] of the 
JDBC thin driver. The Calcite-based engine does not return the partition 
information to the driver, so the driver sends the queries of such a connection 
to a random node of the cluster.
+* The `USE INDEX` syntax. The Calcite-based engine provides the `FORCE_INDEX` 
and `NO_INDEX` <<hints,hints>>.

Review Comment:
   Examples and other improvements of the doc itself is out of scope of this PR.
   Here, I want to fix docs where it's matter for "Calcite now default" 



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to