Operational Commands
Ignite supports the following operational commands:
COPY
Copy data from a CSV file into a SQL table.
COPY FROM '/path/to/local/file.csv'
INTO tableName (columnName, columnName, ...) FORMAT CSV [CHARSET '<charset-name>']
Parameters
-
'/path/to/local/file.csv'
- actual path to your CSV file. -
tableName
- name of the table to which the data will be copied. -
columnName
- name of a column corresponding with the columns in the CSV file.
Description
COPY
allows you to copy the content of a file in the local file system to the server and apply its data to a SQL table. Internally, COPY
reads the file content in a binary form into data packets, and sends those packets to the server. Then, the file content is parsed and executed in a streaming mode. Use this mode if you have data dumped to a file.
Note
|
Currently, COPY is only supported via the JDBC driver and can only work with CSV format.
|
Example
COPY
can be executed like so:
COPY FROM '/path/to/local/file.csv' INTO city (
ID, Name, CountryCode, District, Population) FORMAT CSV
In the above command, substitute /path/to/local/file.csv
with the actual path to your CSV file. For instance, you can use city.csv
which is shipped with the latest Ignite.
You can find it in your {IGNITE_HOME}/examples/src/main/resources/sql/
directory.
SET STREAMING
Stream data in bulk from a file into a SQL table.
SET STREAMING [OFF|ON];
Description
Using the SET
command, you can stream data in bulk into a SQL table in your cluster. When streaming is enabled, the JDBC/ODBC driver will pack your commands in batches and send them to the server (Ignite cluster). On the server side, the batch is converted into a stream of cache update commands which are distributed asynchronously between server nodes. Performing this asynchronously increases peak throughput because at any given time all cluster nodes are busy with data loading.
Usage
To stream data into your cluster, prepare a file with the SET STREAMING ON
command followed by INSERT
commands for data that needs to be loaded.
Note
|
Setting 'STREAMING ON' uses DataStreamer which doesn’t guarantee by default data consistency until successfully finished. |
SET STREAMING ON;
INSERT INTO City(ID, Name, CountryCode, District, Population) VALUES (1,'Kabul','AFG','Kabol',1780000);
INSERT INTO City(ID, Name, CountryCode, District, Population) VALUES (2,'Qandahar','AFG','Qandahar',237500);
INSERT INTO City(ID, Name, CountryCode, District, Population) VALUES (3,'Herat','AFG','Herat',186800);
INSERT INTO City(ID, Name, CountryCode, District, Population) VALUES (4,'Mazar-e-Sharif','AFG','Balkh',127800);
INSERT INTO City(ID, Name, CountryCode, District, Population) VALUES (5,'Amsterdam','NLD','Noord-Holland',731200);
-- More INSERT commands --
Note that before executing the above statements, you should have the tables created in the cluster. Run CREATE TABLE
commands, or provide the commands as part of the file that is used for inserting data, before the SET STREAMING ON
command, like so:
CREATE TABLE City (
ID INT(11),
Name CHAR(35),
CountryCode CHAR(3),
District CHAR(20),
Population INT(11),
PRIMARY KEY (ID, CountryCode)
) WITH "template=partitioned, backups=1, affinityKey=CountryCode, CACHE_NAME=City, KEY_TYPE=demo.model.CityKey, VALUE_TYPE=demo.model.City";
SET STREAMING ON;
INSERT INTO City(ID, Name, CountryCode, District, Population) VALUES (1,'Kabul','AFG','Kabol',1780000);
INSERT INTO City(ID, Name, CountryCode, District, Population) VALUES (2,'Qandahar','AFG','Qandahar',237500);
INSERT INTO City(ID, Name, CountryCode, District, Population) VALUES (3,'Herat','AFG','Herat',186800);
INSERT INTO City(ID, Name, CountryCode, District, Population) VALUES (4,'Mazar-e-Sharif','AFG','Balkh',127800);
INSERT INTO City(ID, Name, CountryCode, District, Population) VALUES (5,'Amsterdam','NLD','Noord-Holland',731200);
-- More INSERT commands --
Note
|
Flush All Data to the ClusterWhen you have finished loading data, make sure to close the JDBC/ODBC connection so that all data is flushed to the cluster. |
Known Limitations
While streaming mode allows you to load data much faster than other data loading techniques mentioned in this guide, it has some limitations:
-
Only
INSERT
commands are allowed; any attempt to executeSELECT
or any other DML or DDL command will cause an exception. -
Due to streaming mode’s asynchronous nature, you cannot know update counts for every statement executed; all JDBC/ODBC commands returning update counts will return 0.
Example
As an example, you can use the sample world.sql file that is shipped with the latest Ignite distribution. It can be found in the {IGNITE_HOME}/examples/sql/
directory. You can use the run
command from tools/sqlline[SQLLine, window=_blank], as shown below:
!run /apache_ignite_version/examples/sql/world.sql
After executing the above command and closing the JDBC connection, all data will be loaded into the cluster and ready to be queried.
KILL QUERY
The KILL QUERY
command allows you to cancel a running query. When a query is cancelled with the KILL
command, all
parts of the query running on all other nodes are terminated as well.
KILL QUERY {ASYNC} 'query_id'
QueryMXBean mxBean = ...;
mxBean.cancelSQL(queryId);
./control.sh --kill SQL query_id
control.bat --kill SQL query_id
Parameters
-
query_id
- can be retrived via the SQL_QUERIES view. -
ASYNC
- is an optional parameter that returns control immediately without waiting for the cancellation to finish.
KILL TRANSACTION
The KILL TRANSACTION
command allows you to cancel a running transaction.
KILL TRANSACTION 'xid'
TransactionMXBean mxBean = ...;
mxBean.cancel(xid);
./control.sh --kill TRANSACTION xid
control.bat --kill TRANSACTION xid
Parameters
-
xid
- the transaction id that can be retrived via the TRANSACTIONS view.
KILL SCAN
The KILL SCAN
command allows you to cancel a running scan query.
KILL SCAN 'origin_node_id' 'cache_name' query_id
QueryMXBean mxBean = ....;
mxBean.cancelScan(originNodeId, cacheName, queryId);
./control.sh --kill SCAN origin_node_id cache_name query_id
control.bat --kill SCAN origin_node_id cache_name query_id
Parameters
-
origin_node_id
,cache_name
,query_id
- can be retrived via the SCAN_QUERIES view.
Example
KILL SCAN '6fa749ee-7cf8-4635-be10-36a1c75267a7_54321' 'cache-name' 1
KILL COMPUTE
KILL CONTINUOUS
The KILL CONTINUOUS
command allows you to cancel a running continuous query.
KILL CONTINUOUS 'origin_node_id', 'routine_id'
QueryMXBean mxBean = ...;
mxBean.cancelContinuous(originNodeId, routineId);
./control.sh --kill CONTINUOUS origin_node_id routine_id
control.bat --kill CONTINUOUS origin_node_id routine_id
Parameters
-
origin_node_id
androutine_id
- can be retrived via the CONTINUOUS_QUERIES view.
Example
KILL CONTINUOUS '6fa749ee-7cf8-4635-be10-36a1c75267a7_54321' '6fa749ee-7cf8-4635-be10-36a1c75267a7_12345'
KILL SERVICE
The KILL SERVICE
command allows you to cancel a running service.
KILL SERVICE 'name'
ServiceMXBean mxBean = ...;
mxBean.cancel(name);
./control.sh --kill SERVICE name
control.bat --kill SERVICE name
Parameters
-
name
- corresponds to the name you selected for the service upon the deployment time. You can always find it with the SERVICES view.
KILL CONSISTENCY repair/check operations
The KILL CONSISTENCY
command allows you to cancel all running consistency repair/check operations.
./control.sh --kill CONSISTENCY
control.bat --kill CONSISTENCY
KILL CLIENT operations
The KILL CLIENT
command allows you to cancel local client (thin/jdbc/odbc) connection.
./control.sh --kill CLIENT connection_id [--node-id node_id]
control.bat --kill CLIENT connection_id [--node-id node_id]
Parameters
-
connection_id
- corresponds to the connection id of the client. Specify ALL to drop all connections. Note, connection_id are local value. Value differs on nodes for the same client. You can always find parameter values with the CLIENT_CONNECTIONS view. -
node_id
- corresponds to the node_id to drop connection from.
Apache, Apache Ignite, the Apache feather and the Apache Ignite logo are either registered trademarks or trademarks of The Apache Software Foundation.