Status

Current state: [One of "Under Discussion", "Accepted", "Rejected"]

Discussion thread: here (<- link to https://mail-archives.apache.org/mod_mbox/flink-dev/)

JIRA: here (<- link to https://issues.apache.org/jira/browse/FLINK-XXXX)

Released: <Flink Version>

Please keep the discussion on the mailing list rather than commenting on the wiki (wiki discussions get unwieldy fast).

Motivation

Often users need to adjust

Public Interfaces

Briefly list any new interfaces that will be introduced as part of this proposal or any existing interfaces that will be removed or changed. The purpose of this section is to concisely call out the public contract that will come along with this feature.

A public interface is any change to the following:

• Binary log format

• The network protocol and api behavior

• Any class in the public packages under clientsConfiguration, especially client configuration

  • org/apache/kafka/common/serialization

  • org/apache/kafka/common

  • org/apache/kafka/common/errors

  • org/apache/kafka/clients/producer

  • org/apache/kafka/clients/consumer (eventually, once stable)

• Monitoring

• Command line tools and arguments

• Anything else that will likely break existing users in some way when they upgrade

Proposed Changes

I suggest introducing a LIKE clause with a following syntax:

CREATE [ TEMPORARY ] TABLE [ IF NOT EXISTS ] table_name ( [
  {   column
    | table_constraint
    | LIKE parent_table [ like_options ] }
    [, ... ]
] )

where like_options are:

  { INCLUDING | EXCLUDING } ALL
| [{ 
     { INCLUDING | EXCLUDING } { CONSTRAINTS }
   | { INCLUDING | EXCLUDING | OVERWRITING } { GENERATED | OPTIONS } }
   [, ...]
  ]

Example:

CREATE [TEMPORARY] TABLE base_table (
    id BIGINT,
    name STRING,
    tstmp TIMESTAMP,
	PRIMARY KEY(id)
) WITH (
    ‘connector.type’: ‘kafka’
)

CREATE [TEMPORARY] TABLE derived_table (
    LIKE base_table,
    WATERMARK FOR tstmp AS tsmp - INTERVAL '5' SECOND
)

Resulting table equivalent to:

CREATE [TEMPORARY] TABLE derived_table (
    id BIGINT,
    name STRING,
    tstmp TIMESTAMP,
	PRIMARY KEY(id),
	WATERMARK FOR tstmp AS tsmp - INTERVAL '5' SECOND
) WITH (
    ‘connector’: ‘kafka’
)

Configuring behaviour of LIKE:

SQL standard defines a way to configure the behaviour of LIKE clause. SQL standards describes options:

• INCLUDING IDENTITY | EXCLUDING IDENTITY
• INCLUDING DEFAULTS | EXCLUDING DEFAULTS
• INCLUDING GENERATED | EXCLUDING GENERATED

PostgresSQL additionally allows for 

CONSTRAINTS | INDEXES | STORAGE | COMMENTS | ALL

I suggest supporting INCLUDING/EXCLUDING:

• ALL
• CONSTRAINTS
• GENERATED
• OPTIONS

Additionally I suggest supporting different INCLUDING/EXCLUDING strategies:

• INCLUDING (default -> fail on duplicate keys)
• EXCLUDING
• OVERWRITING (any explicit options/columns overwrite options/columns from base source, overwriting happens in the order the LIKE clauses were declared)
  

The supported combinations would be:

	INCLUDING	EXCLUDING	OVERWRITING
ALL
CONSTRAINTS
GENERATED
OPTIONS

The reason why I suggest not to support OVERWRITING CONSTRAINTS is that usually it's rather hard to overwrite only some of the constraints. If the constraint(primary key, unique key) from the base table does not apply to the derived table, in my opinion it is safer to exclude all constraints and redefine them.  We could revisit that in the future, if we see a need to support also OVERWRITING for constraints.

Default options:

    INCLUDING ALL = INCLUDING CONSTRAINTS

                                       INCLUDING GENERATED

                                       INCLUDING OPTIONS

It is different from the SQL standard. Nevertheless I think it is worth the change as the primary use case for the clause will be slight adjustments to the original schema rather than rewriting the table completely.

Example:

CREATE [TEMPORARY] TABLE base_table_1 (
    id BIGINT,
    name STRING,
	PRIMARY KEY(id)
) WITH (
    ‘connector’: ‘kafka’,
    ‘connector.starting-offset’: ‘12345’,
    ‘format’: ‘json’
)

CREATE [TEMPORARY] TABLE base_table_2 (
    tstmp TIMESTAMP,
	PRIMARY KEY(tstmp)
) WITH (
    ‘connector’: ‘filesystem’,
    ‘format’: ‘csv’,
    ‘format.delimiter’: ‘\t’   
)

CREATE [TEMPORARY] TABLE derived_table (
    LIKE base_table_1 (
		OVERWRITING SOURCE_OPTIONS),
    LIKE base_table_2 (
		EXCLUDING SOURCE_OPTIONS, 
		EXCLUDING CONSTRAINTS),
    WATERMARK FOR tstmp AS tsmp - INTERVAL '5' SECOND
) WITH (
    ‘connector.starting-offset’: ‘0’
)

The derived_table will be equivalent to:

CREATE [TEMPORARY] TABLE derived_table (
    id BIGINT,
    name STRING,
	tstmp TIMESTAMP,
    WATERMARK FOR tstmp AS tsmp - INTERVAL '5' SECOND,
    PRIMARY KEY(id)
) WITH (
    ‘connector’: ‘kafka’,
    ‘connector.starting-offset’: ‘0’,
    ‘format’: ‘json’
)

Compatibility, Deprecation, and Migration Plan

• What impact (if any) will there be on existing users?
• If we are changing behavior how will we phase out the older behavior?
• If we need special migration tools, describe them here.
• When will we remove the existing behavior?

Test Plan

Describe in few sentences how the FLIP will be tested. We are mostly interested in system tests (since unit-tests are specific to implementation details). How will we know that the implementation works as expected? How will we know nothing broke?

Rejected Alternatives

If there are alternative ways of accomplishing the same thing, what were they? The purpose of this section is to motivate why the design is the way it is and not some other way.