Important

You are viewing documentation for an older version of Confluent Platform. For the latest, click here.

KSQL Syntax Reference

KSQL has similar semantics to SQL:

  • Terminate KSQL statements with a semicolon ;
  • Use a back-slash \ to indicate continuation of a multi-line statement on the next line

Terminology

When using KSQL, the following terminology is used.

Stream

A stream is an unbounded sequence of structured data (“facts”). For example, we could have a stream of financial transactions such as “Alice sent $100 to Bob, then Charlie sent $50 to Bob”. Facts in a stream are immutable, which means new facts can be inserted to a stream, but existing facts can never be updated or deleted. Streams can be created from an Apache Kafka® topic or derived from an existing stream. A stream’s underlying data is durably stored (persisted) within a Kafka topic on the Kafka brokers.

Table

A table is a view of a stream, or another table, and represents a collection of evolving facts. For example, we could have a table that contains the latest financial information such as “Bob’s current account balance is $150”. It is the equivalent of a traditional database table but enriched by streaming semantics such as windowing. Facts in a table are mutable, which means new facts can be inserted to the table, and existing facts can be updated or deleted. Tables can be created from a Kafka topic or derived from existing streams and tables. In both cases, a table’s underlying data is durably stored (persisted) within a Kafka topic on the Kafka brokers.

KSQL CLI Commands

The KSQL CLI commands can be run after starting the KSQL CLI. You can view the KSQL CLI help by running <path-to-confluent>/bin/ksql --help.

Tip: You can search and browse your command history in the KSQL CLI with Ctrl-R. After pressing Ctrl-R, start typing the command or any part of the command to show an auto-complete of past commands.

NAME
        ksql - KSQL CLI

SYNOPSIS
        ksql [ --config-file <configFile> ] [ {-h | --help} ]
                [ --output <outputFormat> ]
                [ --query-row-limit <streamedQueryRowLimit> ]
                [ --query-timeout <streamedQueryTimeoutMs> ] [--] <server>

OPTIONS
        --config-file <configFile>
            A file specifying configs for Ksql and its underlying Kafka Streams
            instance(s). Refer to KSQL documentation for a list of available
            configs.

        -h, --help
            Display help information

        --output <outputFormat>
            The output format to use (either 'JSON' or 'TABULAR'; can be changed
            during REPL as well; defaults to TABULAR)

        --query-row-limit <streamedQueryRowLimit>
            An optional maximum number of rows to read from streamed queries

            This options value must fall in the following range: value >= 1


        --query-timeout <streamedQueryTimeoutMs>
            An optional time limit (in milliseconds) for streamed queries

            This options value must fall in the following range: value >= 1


        --
            This option can be used to separate command-line options from the
            list of arguments (useful when arguments might be mistaken for
            command-line options)

        <server>
            The address of the Ksql server to connect to (ex:
            http://confluent.io:9098)

            This option may occur a maximum of 1 times

RUN SCRIPT

You can run a list of predefined queries and commands from in a file by using the RUN SCRIPT command.

Example:

RUN SCRIPT '/local/path/to/queries.sql';

The RUN SCRIPT command supports a subset of KSQL statements:

It does not support statements such as:

  • SHOW TOPICS and SHOW STREAMS etc
  • TERMINATE
  • Non-persistent queries: SELECT etc

RUN SCRIPT can also be used from the command line, for instance when writing shell scripts. For more information, see Running KSQL Statements From the Command Line.

KSQL statements

Tip

  • KSQL statements must be terminated with a semicolon (;).
  • Multi-line statements:
    • In the CLI you must use a backslash (\) to indicate continuation of a statement on the next line.
    • Do not use \ for multi-line statements in .sql files.
  • The hyphen character, -, isn’t supported in names for streams, tables, topics, and columns.
  • Don’t use quotes around stream names or table names when you CREATE them.

CREATE STREAM

Synopsis

CREATE STREAM stream_name ( { column_name data_type } [, ...] )
  WITH ( property_name = expression [, ...] );

Description

Create a new stream with the specified columns and properties.

The supported column data types are:

  • BOOLEAN
  • INTEGER
  • BIGINT
  • DOUBLE
  • VARCHAR (or STRING)
  • ARRAY<ArrayType> (JSON and AVRO only)
  • MAP<VARCHAR, ValueType> (JSON and AVRO only)

KSQL adds the implicit columns ROWTIME and ROWKEY to every stream and table, which represent the corresponding Kafka message timestamp and message key, respectively.

The WITH clause supports the following properties:

Property Description
KAFKA_TOPIC (required) The name of the Kafka topic that backs this stream. The topic must already exist in Kafka.
VALUE_FORMAT (required) Specifies the serialization format of the message value in the topic. Supported formats: JSON, DELIMITED (comma-separated value), and AVRO.
KEY Optimization hint: If the Kafka message key is also present as a field/column in the Kafka message value, you may set this property to associate the corresponding field/column with the implicit ROWKEY column (message key). If set, KSQL uses it as an optimization hint to determine if repartitioning can be avoided when performing aggregations and joins. You can only use this if the key format in kafka is VARCHAR or STRING. Do not use this hint if the message key format in kafka is AVRO or JSON. See Key Requirements for more information.
TIMESTAMP By default, the implicit ROWTIME column is the timestamp of the message in the Kafka topic. The TIMESTAMP property can be used to override ROWTIME with the contents of the specified field/column within the Kafka message value (similar to timestamp extractors in Kafka’s Streams API). Time-based operations such as windowing will process a record according to the timestamp in ROWTIME.

Note

  • To use Avro, you must have Schema Registry enabled and ksql.schema.registry.url must be set in the KSQL server configuration file. See Avro and Schema Registry.
  • Avro field names are not case sensitive in KSQL. This matches the KSQL column name behavior.

Example:

CREATE STREAM pageviews (viewtime BIGINT, user_id VARCHAR, page_id VARCHAR)
  WITH (VALUE_FORMAT = 'JSON',
        KAFKA_TOPIC = 'my-pageviews-topic');

CREATE TABLE

Synopsis

CREATE TABLE table_name ( { column_name data_type } [, ...] )
  WITH ( property_name = expression [, ...] );

Description

Create a new table with the specified columns and properties.

The supported column data types are:

  • BOOLEAN
  • INTEGER
  • BIGINT
  • DOUBLE
  • VARCHAR (or STRING)
  • ARRAY<ArrayType> (JSON and AVRO only)
  • MAP<VARCHAR, ValueType> (JSON and AVRO only)

KSQL adds the implicit columns ROWTIME and ROWKEY to every stream and table, which represent the corresponding Kafka message timestamp and message key, respectively.

KSQL has currently the following requirements for creating a table from a Kafka topic:

  1. The Kafka message key must also be present as a field/column in the Kafka message value. The KEY property (see below) must be defined to inform KSQL which field/column in the message value represents the key. If the message key is not present in the message value, follow the instructions in Key Requirements.
  2. The message key must be in VARCHAR aka STRING format. If the message key is not in this format, follow the instructions in Key Requirements.

The WITH clause supports the following properties:

Property Description
KAFKA_TOPIC (required) The name of the Kafka topic that backs this table. The topic must already exist in Kafka.
VALUE_FORMAT (required) Specifies the serialization format of message values in the topic. Supported formats: JSON, DELIMITED (comma-separated value), and AVRO.
KEY (required)

Associates a field/column within the Kafka message value with the implicit ROWKEY column (message key) in the KSQL table.

KSQL currently requires that the Kafka message key, which will be available as the implicit ROWKEY column in the table, must also be present as a field/column in the message value. You must set the KEY property to this corresponding field/column in the message value, and this column must be in VARCHAR aka STRING format. See Key Requirements for more information.
TIMESTAMP By default, the implicit ROWTIME column is the timestamp of the message in the Kafka topic. The TIMESTAMP property can be used to override ROWTIME with the contents of the specified field/column within the Kafka message value (similar to timestamp extractors in Kafka’s Streams API). Time-based operations such as windowing will process a record according to the timestamp in ROWTIME.

Note

  • To use Avro, you must have Schema Registry enabled and ksql.schema.registry.url must be set in the KSQL server configuration file. See Avro and Schema Registry.
  • Avro field names are not case sensitive in KSQL. This matches the KSQL column name behavior.

Example:

CREATE TABLE users (usertimestamp BIGINT, user_id VARCHAR, gender VARCHAR, region_id VARCHAR) WITH (
    KAFKA_TOPIC = 'my-users-topic',
    KEY = 'user_id');

CREATE STREAM AS SELECT

Synopsis

CREATE STREAM stream_name
  [WITH ( property_name = expression [, ...] )]
  AS SELECT  select_expr [, ...]
  FROM from_item
  [ LEFT JOIN join_table ON join_criteria ]
  [ WHERE condition ]
  [PARTITION BY column_name];

Description

Create a new stream along with the corresponding Kafka topic, and continuously write the result of the SELECT query into the stream and its corresponding topic.

If the PARTITION BY clause is present, then the resulting stream will have the specified column as its key.

The WITH clause for the result supports the following properties:

Property Description
KAFKA_TOPIC The name of the Kafka topic that backs this stream. If this property is not set, then the name of the stream in upper case will be used as default.
VALUE_FORMAT Specifies the serialization format of the message value in the topic. Supported formats: JSON, DELIMITED (comma-separated value), and AVRO. If this property is not set, then the format of the input stream/table is used.
PARTITIONS The number of partitions in the backing topic. If this property is not set, then the number of partitions is taken from the value of the ksql.sink.partitions property, which defaults to four partitions. The ksql.sink.partitions property can be set in the properties file the KSQL server is started with, or by using the SET statement.
REPLICAS The replication factor for the topic. If this property is not set, then the number of replicas of the input stream or table will be used.
TIMESTAMP

Sets a field within this stream’s schema to be used as the default source of ROWTIME for any downstream queries. Downstream queries that use time-based operations, such as windowing, will process records in this stream based on the timestamp in this field. By default, such queries will also use this field to set the timestamp on any records emitted to Kafka.

If not supplied, the ROWTIME of the source stream will be used.

Note: This doesn’t affect the processing of the query that populates this stream. For example, given the following statement:

The window into which each row of bar is placed is determined by bar’s ROWTIME, not t2.

Note

  • To use Avro, you must have Schema Registry enabled and ksql.schema.registry.url must be set in the KSQL server configuration file. See Avro and Schema Registry.
  • Avro field names are not case sensitive in KSQL. This matches the KSQL column name behavior.

Note: The KEY property is not supported – use PARTITION BY instead.

CREATE TABLE AS SELECT

Synopsis

CREATE TABLE table_name
  [WITH ( property_name = expression [, ...] )]
  AS SELECT  select_expr [, ...]
  FROM from_item
  [ WINDOW window_expression ]
  [ WHERE condition ]
  [ GROUP BY grouping_expression ]
  [ HAVING having_expression ];

Description

Create a new KSQL table along with the corresponding Kafka topic and stream the result of the SELECT query as a changelog into the topic. Note that WINDOW, GROUP BY and HAVING clauses can only be used if the from_item is a stream.

The WITH clause supports the following properties:

Property Description
KAFKA_TOPIC The name of the Kafka topic that backs this table. If this property is not set, then the name of the table will be used as default.
VALUE_FORMAT Specifies the serialization format of the message value in the topic. Supported formats: JSON, DELIMITED (comma-separated value), and AVRO. If this property is not set, then the format of the input stream or table is used.
PARTITIONS The number of partitions in the backing topic. If this property is not set, then the number of partitions is taken from the value of the ksql.sink.partitions property, which defaults to four partitions. The ksql.sink.partitions property can be set in the properties file the KSQL server is started with, or by using the SET statement.
REPLICAS The replication factor for the topic. If this property is not set, then the number of replicas of the input stream or table will be used.
TIMESTAMP

Sets a field within this tables’s schema to be used as the default source of ROWTIME for any downstream queries. Downstream queries that use time-based operations, such as windowing, will process records in this stream based on the timestamp in this field.

If not supplied, the ROWTIME of the source stream will be used.

Note: This doesn’t affect the processing of the query that populates this table. For example, given the following statement:

The window into which each row of bar is placed is determined by bar’s ROWTIME, not t2.

Note

  • To use Avro, you must have Schema Registry enabled and ksql.schema.registry.url must be set in the KSQL server configuration file. See Avro and Schema Registry.
  • Avro field names are not case sensitive in KSQL. This matches the KSQL column name behavior.

DESCRIBE

Synopsis

DESCRIBE [EXTENDED] (stream_name|table_name);

Description

  • DESCRIBE: List the columns in a stream or table along with their data type and other attributes.
  • DESCRIBE EXTENDED: Display DESCRIBE information with additional runtime statistics, Kafka topic details, and the set of queries that populate the table or stream.

Example of describing a table:

DESCRIBE ip_sum;

Your output should resemble:

 Field   | Type
-------------------------------------
 ROWTIME | BIGINT           (system)
 ROWKEY  | VARCHAR(STRING)  (system)
 IP      | VARCHAR(STRING)  (key)
 KBYTES  | BIGINT
-------------------------------------
For runtime statistics and query details run: DESCRIBE EXTENDED <Stream,Table>

Example of describing a table with extended information:

DESCRIBE EXTENDED ip_sum;

Your output should resemble:

Type                 : TABLE
Key field            : CLICKSTREAM.IP
Timestamp field      : Not set - using <ROWTIME>
Key format           : STRING
Value format         : JSON
Kafka output topic   : IP_SUM (partitions: 4, replication: 1)

 Field   | Type
-------------------------------------
 ROWTIME | BIGINT           (system)
 ROWKEY  | VARCHAR(STRING)  (system)
 IP      | VARCHAR(STRING)  (key)
 KBYTES  | BIGINT
-------------------------------------

Queries that write into this TABLE
-----------------------------------
id:CTAS_IP_SUM - CREATE TABLE IP_SUM as SELECT ip,  sum(bytes)/1024 as kbytes FROM CLICKSTREAM window SESSION (300 second) GROUP BY ip;

For query topology and execution plan run: EXPLAIN <QueryId>; for more information

Local runtime statistics
------------------------
messages-per-sec:      4.41   total-messages:       486     last-message: 12/14/17 4:32:23 PM GMT
 failed-messages:         0      last-failed:       n/a
(Statistics of the local KSQL server interaction with the Kafka topic IP_SUM)

EXPLAIN

Synopsis

EXPLAIN (sql_expression|query_id);

Description

Show the execution plan for a SQL expression or, given the ID of a running query, show the execution plan plus additional runtime information and metrics. Statements such as DESCRIBE EXTENDED, for example, show the IDs of queries related to a stream or table.

Example of explaining a running query:

EXPLAIN ctas_ip_sum;

Your output should resemble:

Type                 : QUERY
SQL                  : CREATE TABLE IP_SUM as SELECT ip,  sum(bytes)/1024 as kbytes FROM CLICKSTREAM window SESSION (300 second) GROUP BY ip;


Local runtime statistics
------------------------
messages-per-sec:     104.38   total-messages:       14238     last-message: 12/14/17 4:30:42 PM GMT
 failed-messages:          0      last-failed:         n/a
(Statistics of the local Ksql Server interaction with the Kafka topic IP_SUM)

Execution plan
--------------
 > [ PROJECT ] Schema: [IP : STRING , KBYTES : INT64].
         > [ AGGREGATE ] Schema: [CLICKSTREAM.IP : STRING , CLICKSTREAM.BYTES : INT64 , KSQL_AGG_VARIABLE_0 : INT64].
                 > [ PROJECT ] Schema: [CLICKSTREAM.IP : STRING , CLICKSTREAM.BYTES : INT64].
                         > [ REKEY ] Schema: [CLICKSTREAM.ROWTIME : INT64 , CLICKSTREAM.ROWKEY : STRING , CLICKSTREAM._TIME : INT64 , CLICKSTREAM.TIME : STRING , CLICKSTREAM.IP : STRING , CLICKSTREAM.REQUEST : STRING , CLICKSTREAM.STATUS : INT32 , CLICKSTREAM.USERID : INT32 , CLICKSTREAM.BYTES : INT64 , CLICKSTREAM.AGENT : STRING].
                                 > [ SOURCE ] Schema: [CLICKSTREAM.ROWTIME : INT64 , CLICKSTREAM.ROWKEY : STRING , CLICKSTREAM._TIME : INT64 , CLICKSTREAM.TIME : STRING , CLICKSTREAM.IP : STRING , CLICKSTREAM.REQUEST : STRING , CLICKSTREAM.STATUS : INT32 , CLICKSTREAM.USERID : INT32 , CLICKSTREAM.BYTES : INT64 , CLICKSTREAM.AGENT : STRING].


Processing topology
-------------------
Sub-topologies:
  Sub-topology: 0
    Source: KSTREAM-SOURCE-0000000000 (topics: [clickstream])
      --> KSTREAM-MAP-0000000001
    Processor: KSTREAM-MAP-0000000001 (stores: [])
      --> KSTREAM-TRANSFORMVALUES-0000000002
      <-- KSTREAM-SOURCE-0000000000

DROP STREAM

Synopsis

DROP STREAM stream_name;

Description

Drops an existing stream.

DROP TABLE

Synopsis

DROP TABLE table_name;

Description

Drops an existing table.

PRINT

PRINT qualifiedName [FROM BEGINNING] [INTERVAL]

Description

Print the contents of Kafka topics to the KSQL CLI.

Important

SQL grammar defaults to uppercase formatting. You can use quotations (") to print topics that contain lowercase characters.

The PRINT statement supports the following properties:

Property Description
FROM BEGINNING Print starting with the first message in the topic. If not specified, PRINT starts with the most recent message.
INTERVAL Print every nth message. The default is 1, meaning that every message is printed.

For example:

PRINT 'ksql__commands' FROM BEGINNING;

Your output should resemble:

Format:JSON
{"ROWTIME":1516010696273,"ROWKEY":"\"stream/CLICKSTREAM/create\"","statement":"CREATE STREAM clickstream (_time bigint,time varchar, ip varchar, request varchar, status int, userid int, bytes bigint, agent varchar) with (kafka_topic = 'clickstream', value_format = 'json');","streamsProperties":{}}
{"ROWTIME":1516010709492,"ROWKEY":"\"table/EVENTS_PER_MIN/create\"","statement":"create table events_per_min as select userid, count(*) as events from clickstream window  TUMBLING (size 10 second) group by userid;","streamsProperties":{}}
^CTopic printing ceased

SELECT

Synopsis

SELECT select_expr [, ...]
  FROM from_item
  [ LEFT JOIN join_table ON join_criteria ]
  [ WINDOW window_expression ]
  [ WHERE condition ]
  [ GROUP BY grouping_expression ]
  [ HAVING having_expression ];

Description

Selects rows from a KSQL stream or table. The result of this statement will not be persisted in a Kafka topic and will only be printed out in the console. To stop the continuous query in the CLI press Ctrl-C. Note that WINDOW, GROUP BY and HAVING clauses can only be used if the from_item is a stream.

In the above statements from_item is one of the following:

  • stream_name [ alias ]
  • table_name [ alias ]
  • from_item LEFT JOIN from_item ON join_condition

The WHERE clause can refer to any column defined for a stream or table, including the two implicit columns ROWTIME and ROWKEY.

Example:

SELECT * FROM pageviews
  WHERE ROWTIME >= 1510923225000
    AND ROWTIME <= 1510923228000;

Tip: If you want to select older data, you can configure KSQL to query the stream from the beginning. You must run this configuration before running the query:

SET 'auto.offset.reset' = 'earliest';

The WINDOW clause lets you control how to group input records that have the same key into so-called windows for operations such as aggregations or joins. Windows are tracked per record key. KSQL supports the following WINDOW types:

  • TUMBLING: Tumbling windows group input records into fixed-sized, non-overlapping windows based on the records’ timestamps. You must specify the window size for tumbling windows. Note: Tumbling windows are a special case of hopping windows where the window size is equal to the advance interval.

    Example:

    SELECT item_id, SUM(quantity)
  FROM orders
  WINDOW TUMBLING (SIZE 20 SECONDS)
  GROUP BY item_id;

  • HOPPING: Hopping windows group input records into fixed-sized, (possibly) overlapping windows based on the records’ timestamps. You must specify the window size and the advance interval for hopping windows.

    Example:

    SELECT item_id, SUM(quantity)
  FROM orders
  WINDOW HOPPING (SIZE 20 SECONDS, ADVANCE BY 5 SECONDS)
  GROUP BY item_id;

  • SESSION: Session windows group input records into so-called sessions. You must specify the session inactivity gap parameter for session windows. For example, imagine you set the inactivity gap to 5 minutes. If, for a given record key such as “alice”, no new input data arrives for more than 5 minutes, then the current session for “alice” is closed, and any newly arriving data for “alice” in the future will mark the beginning of a new session.

    Example:

    SELECT item_id, SUM(quantity)
  FROM orders
  WINDOW SESSION (20 SECONDS)
  GROUP BY item_id;

The following list shows valid time units for SIZE, ADVANCE BY, and SESSION.

  • DAY, DAYS
  • HOUR, HOURS
  • MINUTE, MINUTES
  • SECOND, SECONDS
  • MILLISECOND, MILLISECONDS

For more information, see windows_in_ksql_queries.

CAST

Synopsis

CAST (expression AS data_type);

You can cast an expression’s type to a new type using CAST. Here is an example of converting a BIGINT into a VARCHAR type:

-- This query converts the numerical count into a suffixed string; e.g., 5 becomes '5_HELLO'
SELECT page_id, CONCAT(CAST(COUNT(*) AS VARCHAR), '_HELLO')
  FROM pageviews_enriched
  WINDOW TUMBLING (SIZE 20 SECONDS)
  GROUP BY page_id;

LIKE

Synopsis

column_name LIKE pattern;

The LIKE operator is used for prefix or suffix matching. Currently KSQL supports %, which represents zero or more characters.

Example:

SELECT user_id
  FROM users
  WHERE user_id LIKE 'santa%';

SHOW TOPICS

Synopsis

SHOW | LIST TOPICS;

Description

List the available topics in the Kafka cluster that KSQL is configured to connect to (default setting for bootstrap.servers: localhost:9092).

SHOW STREAMS

Synopsis

SHOW | LIST STREAMS;

Description

List the defined streams.

SHOW TABLES

Synopsis

SHOW | LIST TABLES;

Description

List the defined tables.

SHOW QUERIES

Synopsis

SHOW QUERIES;

Description

List the running persistent queries.

SHOW PROPERTIES

Synopsis

SHOW PROPERTIES;

Description

List the configuration settings that are currently in effect.

TERMINATE

Synopsis

TERMINATE query_id;

Description

Terminate a persistent query. Persistent queries run continuously until they are explicitly terminated.

  • In client-server mode, exiting the CLI will not stop persistent queries because the KSQL server(s) will continue to process the queries.

(To terminate a non-persistent query use Ctrl-C in the CLI.)

Scalar functions

Function Example Description
ABS ABS(col1) The absolute value of a value
ARRAYCONTAINS ARRAYCONTAINS('[1, 2, 3]', 3) Given JSON or AVRO array checks if a search value contains in it
CEIL CEIL(col1) The ceiling of a value
CONCAT CONCAT(col1, '_hello') Concatenate two strings
EXTRACTJSONFIELD EXTRACTJSONFIELD(message, '$.log.cloud') Given a string column in JSON format, extract the field that matches
FLOOR FLOOR(col1) The floor of a value
IFNULL IFNULL(col1, retval) If the provided VARCHAR is NULL, return retval, otherwise, return the value. Only VARCHAR values are supported for the input. The return value must be a VARCHAR.
LCASE LCASE(col1) Convert a string to lowercase
LEN LEN(col1) The length of a string
RANDOM RANDOM() Return a random DOUBLE value between 0.0 and 1.0
ROUND ROUND(col1) Round a value to the nearest BIGINT value
STRINGTOTIMESTAMP STRINGTOTIMESTAMP(col1, 'yyyy-MM-dd HH:mm:ss.SSS') Converts a string value in the given format into the BIGINT value that represents the timestamp.
SUBSTRING SUBSTRING(col1, 2, 5) Return the substring with the start and end indices
TIMESTAMPTOSTRING TIMESTAMPTOSTRING(ROWTIME, 'yyyy-MM-dd HH:mm:ss.SSS') Converts a BIGINT timestamp value into the string representation of the timestamp in the given format.
TRIM TRIM(col1) Trim the spaces from the beginning and end of a string
UCASE UCASE(col1) Convert a string to uppercase

Aggregate functions

Function Example Description
COUNT COUNT(col1) Count the number of rows
MAX MAX(col1) Return the maximum value for a given column and window
MIN MIN(col1) Return the minimum value for a given column and window
SUM SUM(col1) Sums the column values
TOPK TOPK(col1, k) Return the Top K values for the given column and window
TOPKDISTINCT TOPKDISTINCT(col1, k) Return the distinct Top K values for the given column and window

Key Requirements

Message Keys

The CREATE STREAM and CREATE TABLE statements, which read data from a Kafka topic into a stream or table, allow you to specify a field/column in the Kafka message value that corresponds to the Kafka message key by setting the KEY property of the WITH clause.

Example:

CREATE TABLE users (registertime BIGINT, gender VARCHAR, regionid VARCHAR, userid VARCHAR)
  WITH (KAFKA_TOPIC='users', VALUE_FORMAT='JSON', KEY = 'userid');

The KEY property is:

  • Required for tables.

  • Optional for streams. Here, KSQL uses it as an optimization hint to determine if repartitioning can be avoided when performing aggregations and joins.

    Important

    Don’t set the KEY property, unless you have validated that your stream doesn’t need to be re-partitioned for future joins. If you set the KEY property, you will need to re-partition explicitly if your record key doesn’t meet partitioning requirements.

In either case, when setting KEY you must be sure that both of the following conditions are true:

  1. For every record, the contents of the Kafka message key must be the same as the contents of the column set in KEY (which is derived from a field in the Kafka message value).
  2. KEY must be set to a column of type VARCHAR aka STRING.

If these conditions are not met, then the results of aggregations and joins may be incorrect. However, if your data doesn’t meet these requirements, you can still use KSQL with a few extra steps. The following section explains how.

Table-table joins can be joined only on the KEY field, and one-to-many (1:N) joins aren’t supported.

What To Do If Your Key Is Not Set or Is In A Different Format

Streams

For streams, just leave out the KEY property from the WITH clause. KSQL will take care of repartitioning the stream for you using the value(s) from the GROUP BY columns for aggregates, and the join predicate for joins.

Tables

For tables, you can still use KSQL if the message key is not also present in the Kafka message value or if it is not in the required format as long as one of the following statements is true:

  • The message key is a unary function of the value in the desired key column.
  • It is ok for the messages in the topic to be re-ordered before being inserted into the table.

First create a stream to have KSQL write the message key, and then declare the table on the output topic of this stream:

Example:

  • Goal: You want to create a table from a topic, which is keyed by userid of type INT.
  • Problem: The message key is present as a field/column (aptly named userid) in the message value but in the wrong format (INT instead of VARCHAR).
-- Create a stream on the original topic
CREATE STREAM users_with_wrong_key_format (userid INT, username VARCHAR, email VARCHAR)
  WITH (KAFKA_TOPIC='users', VALUE_FORMAT='JSON');

-- Derive a new stream with the required key changes.
-- 1) The CAST statement converts the key to the required format.
-- 2) The PARTITION BY clause re-partitions the stream based on the new, converted key.
CREATE STREAM users_with_proper_key
  WITH(KAFKA_TOPIC='users-with-proper-key') AS
  SELECT CAST(userid as VARCHAR) as userid_string, username, email
  FROM users_with_wrong_key_format
  PARTITION BY userid_string;

-- Now you can create the table on the properly keyed stream.
CREATE TABLE users_table (userid_string VARCHAR, username VARCHAR, email VARCHAR)
  WITH (KAFKA_TOPIC='users-with-proper-key',
        VALUE_FORMAT='JSON',
        KEY='userid_string');

Example:

  • Goal: You want to create a table from a topic, which is keyed by userid of type INT.
  • Problem: The message key is not present as a field/column in the topic’s message values.
-- Create a stream on the original topic.
-- The topic is keyed by userid, which is available as the implicit column ROWKEY
-- in the users_with_missing_key stream. Note how the explicit column definitions
-- only define username and email but not userid.
CREATE STREAM users_with_missing_key (username VARCHAR, email VARCHAR)
  WITH (KAFKA_TOPIC='users', VALUE_FORMAT='JSON');

-- Derive a new stream with the required key changes.
-- 1) The contents of ROWKEY (message key) are copied into the message value as the userid_string column,
--    and the CAST statement converts the key to the required format.
-- 2) The PARTITION BY clause re-partitions the stream based on the new, converted key.
CREATE STREAM users_with_proper_key
  WITH(KAFKA_TOPIC='users-with-proper-key') AS
  SELECT CAST(ROWKEY as VARCHAR) as userid_string, username, email
  FROM users_with_missing_key
  PARTITION BY userid_string;

-- Now you can create the table on the properly keyed stream.
CREATE TABLE users_table (userid_string VARCHAR, username VARCHAR, email VARCHAR)
  WITH (KAFKA_TOPIC='users-with-proper-key',
        VALUE_FORMAT='JSON',
        KEY='userid_string');