rabbitmq-v3.yaml (Beta)

GPSS load configuration file for a RabbitMQ data source (version 3).

Synopsis

version: v3

targets:
- gpdb:
    host: <host>
    port: <greenplum_port>
    user: <user_name>
    password: <password>
    database: <db_name>
    work_schema: <work_schema_name>
    error_limit: <num_errors> | <percentage_errors>
    filter_expression: <filter_string>
    tables:
      - table: <table_name>
        schema: <schema_name>
        mode:
          # specify a single mode property block (described below)
          insert: {}
          update:
            <mode_specific_property>: <value>
            ...
          merge:
            <mode_specific_property>: <value>
            ...
        transformer:
          transform: <udf_transform_udf_name>
          properties:
            <udf_transform_property_name>: <property_value>
            ...
          columns:
            - <udf_transform_column_name>
            ...
        mapping:
          <target_column_name> : <source_column_name> | <expression>
          ...
        filter: <output_filter_string>

sources:
- rabbitmq:
    server: <rmq_user>:<rmq_password>@<rmq_host>:<rmq_port>
    vhost: <gpss_vhost>
    stream: <name> | queue: <name>
    consistency: strong | at-least | at-most | none
    fallback_offset: earliest | latest
    save_failing_batch: <boolean>
    recover_failing_batch: <boolean> (Beta)
    data_content:
      <data_format>:
        <column_spec>
        <other_props>
    meta:
      json:
        column:
          name: meta
          type: json
    encoding: <char_set>
    transformer:
      path: <path_to_plugin_transform_library>
      on_init: <plugin_transform_init_name>
      transform: <plugin_transform_name>
      properties:
        <plugin_transform_property_name>: <property_value>
        ...
    properties:
      <rmq_property_name>: <rmq_property_value>
      ...
    task:
      batch_size:
        max_count: <number_of_rows>
        interval_ms: <wait_time>
        idle_duration_ms: <idle_time>
      window_size: <num_batches>
      window_statement: <udf_or_sql_to_run>
      prepare_statement: <udf_or_sql_to_run>
      teardown_statement: <udf_or_sql_to_run>

option:
  schedule:
    max_retries: <num_retries>
    retry_interval: <retry_time>
    running_duration: <run_time>
    auto_stop_restart_interval: <restart_time>
    max_restart_times: <num_restarts>
    quit_at_eof_after: <clock_time>
  alert:
    command: <command_to_run>
    workdir: <directory>
    timeout: <alert_time>

Where the mode_specific_propertys that you can specify for update and merge mode follow:

update:
  match_columns: [<match_column_names>]
  order_columns: [<order_column_names>]
  update_columns: [<update_column_names>]
  update_condition: <update_condition>

merge:
  match_columns: [<match_column_names>]
  update_columns: [<update_column_names>]
  order_columns: [<order_column_names>]
  update_condition: <update_condition>
  delete_condition: <delete_condition>

Where data_format, column_spec, and other_props are one of the following blocks

binary:
  source_column_name: <column_name>

csv:
  columns:
    - name: <column_name>
      type: <column_data_type>
    ...
  delimiter: <delim_char>
  quote: <quote_char>
  null_string: <nullstr_val>
  escape: <escape_char>
  force_not_null: <columns>
  fill_missing_fields: <boolean>

custom:
  columns:
    - name: <column_name>
      type: <column_data_type>
    ...
  name: <formatter_name>
  options:
    - <optname>=<optvalue>
    ...

delimited:
  columns:
    - name: <column_name>
      type: <column_data_type>
    ...
  delimiter: <delimiter_string>
  eol_prefix: <prefix_string>
  quote: <quote_char>
  escape: <escape_char>

json:
  column:
    name: <column_name>
    type: json | jsonb
  is_jsonl: <boolean>
  newline: <newline_str>

And where you may specify any property value with a template variable that GPSS substitutes at runtime using the following syntax:

<property:> {{<template_var>}}

Description

Note
Version 3 of the GPSS load configuration file is different in both content and format than previous versions of the file. Certain symbols used in the GPSS version 1 and 2 configuration file reference page syntax have different meanings in version 3 syntax:

Brackets [] are literal and are used to specify a list in version 3. They are no longer used to signify the optionality of a property.

Curly braces {} are literal and are used to specify YAML mappings in version 3 syntax. They are no longer used with the pipe symbol (|) to identify a list of choices.

You specify load configuration properties for a Greenplum Streaming Server (GPSS) RabbitMQ load job in a YAML-formatted configuration file. (This reference page uses the name rabbitmq-v3.yaml when referring to this file; you may choose your own name for the file.) Load properties include Greenplum Database connection and data import properties, RabbitMQ data source information, and properies specific to the GPSS job.

The gpsscli utilities process the YAML configuration file in order, using indentation (spaces) to determine the document hierarchy and the relationships between the sections. The use of white space in the file is significant. Keywords are not case-sensitive.

Keywords and Values

version Property

version: v3: The version of the configuration file. You must specify version: v3.

targets:gpdb Properties

host: host

The host name or IP address of the Greenplum Database coordinator host.

port: greenplum_port

The port number of the Greenplum Database server on the coordinator host.

user: user_name

The name of the Greenplum Database user/role. This user_name must have permissions as described in the Configuring Greenplum Database Role Privileges.

password: password

The password for the Greenplum Database user/role.

database: db_name

The name of the Greenplum database.

work_schema: work_schema_name

The name of the Greenplum Database schema in which GPSS creates internal tables. The default work_schema_name is public.

error_limit: num_errors | percentage_errors

The error threshold, specified as either an absolute number or a percentage. GPSS stops running the job when this limit is reached.

filter_expression: filter_string

The filter to apply to the input data before GPSS loads the data into Greenplum Database. If the filter evaluates to true, GPSS loads the message. If the filter evaluates to false, the message is dropped. filter_string must be a valid SQL conditional expression and may reference one or more source value or meta column names.

tables:

The Greenplum Database tables, and the data that GPSS will load into each.

Note
GPSS supports loading from a RabbitMQ data source into a single Greenplum Database table only.

table: table_name: The name of the Greenplum Database table into which GPSS loads the data.
schema: schema_name: The name of the Greenplum Database schema in which table_name resides. Optional, the default schema is the public schema.
mode:: The table load mode; insert, merge, or update. The default mode is insert.

Note
update and merge are not supported if the target table column name is a reserved keyword, has capital letters, or includes any character that requires quotes (" ") to identify the column.

insert:: Inserts source data into Greenplum.
update:: Updates the target table columns that are listed in update_columns when the input columns identified in match_columns match the named target table columns and the optional update_condition is true.
merge:: Inserts new rows and updates existing rows when:

columns are listed in update_columns,
the match_columns target table column values are equal to the input data, and
an optional update_condition is specified and met.

Deletes rows when:

the match_columns target table column values are equal to the input data, and
an optional delete_condition is specified and met.

New rows are identified when the match_columns value in the source data does not have a corresponding value in the existing data of the target table. In those cases, the entire row from the source file is inserted, not only the match_columns and update_columns. If there are multiple new match_columns values in the input data that are the same, GPSS inserts or updates the target table using a random matching input row. When you specify order_columns, GPSS sorts the input data on the specified column(s) and inserts or updates from the input row with the largest value.

mode_property_name: value: The name to value mapping for a mode property. Each mode supports one or more of the following properties as specified in the Synopsis.
match_columns: [match_column_names]: A comma-separated list that specifies the column(s) to use as the join condition for the update. The attribute value in the specified target column(s) must be equal to that of the corresponding source data column(s) in order for the row to be updated in the target table.; Required when mode is merge or update.
order_columns: [order_column_names]: A comma-separated list that specifies the column(s) by which GPSS sorts the rows. When multiple matching rows exist in a batch, order_columns is used with match_columns to determine the input row with the largest value; GPSS uses that row to write/update the target.; Optional. May be specified in merge mode to sort the input data rows.
update_columns: [update_column_names]: A column-sparated list that specifies the column(s) to update for the rows that meet the match_columns criteria and the optional update_condition.; Required when mode is merge or update.
update_condition: update_condition: Specifies a boolean condition, similar to that which you would declare in a WHERE clause, that must be met in order for a row in the target table to be updated (or inserted, in the case of a merge). Optional.
delete_condition: delete_condition: In merge mode, specifies a boolean condition, similar to that which you would declare in a WHERE clause, that must be met for GPSS to delete rows in the target table that meet the match_columns criteria. Optional.
transformer:: Optional. Output data transform block. An output data transformer is a user-defined function (UDF) that transforms the data before it is loaded into Greenplum Database. The semantics of the UDF are transform-specific.

Note
GPSS currently supports specifying only one of the mapping or (UDF) transformer blocks in the load configuration file, not both.
transform: udf_transform_udf_name: The name of the output transform UDF. GPSS invokes this function for every batch of data it writes to Greenplum Database.
properties: udf_transform_property_name: property_value: One or more property name and value pairs that GPSS passes to udf_transform_udf_name.
columns: output_transform_column_name: The name of one or more columns involved in the transform.
mapping:: Optional. Overrides the default source-to-target column mapping.

Note
GPSS currently supports specifying only one of the mapping or (UDF) transformer blocks in the load configuration file, not both.

Note
When you specify a mapping, ensure that you provide a mapping for all source data elements of interest. GPSS does not automatically match column names when you provide a mapping block.

target_column_name: source_column_name | expression: target_column_name specifies the target Greenplum Database table column name. GPSS maps this column name to the source column name specified in source_column_name, or to an expression. When you specify an expression, you may provide a value expression that you would specify in the SELECT list of a query, such as a constant value, a column reference, an operator invocation, a built-in or user-defined function call, and so on.
filter: output_filter_string: The filter to apply to the output data before GPSS loads the data into Greenplum Database. If the filter evaluates to true, GPSS loads the message. If the filter evaluates to false, the message is dropped. output_filter_string must be a valid SQL conditional expression and may reference one or more META or VALUE column names.

sources:rabbitmq: Options

server: rmq_user:rmq_password@rmq_host:rmq_port: The RabbitMQ server connection string; includes the user name with which RabbitMQ logs in to the broker, the password for rmq_user, the hostname or IP address of the RabbitMQ server, and the port number on which the RabbitMQ server is listening. rmq_user and rmq_password are optional, and must be omitted when loading from a RabbitMQ queue over a TLS-encrypted connection.
vhost: gpss_vhost: The RabbitMQ virtual host that represents the GPSS server.
stream: name: The name of the RabbitMQ stream from which to read the data. You may specify only one of stream or queue.
queue: name: The name of the RabbitMQ queue from which to read the data. You may specify only one of stream or queue.
consistency: strong | at-least | at-most | none: Specify how GPSS should manage message offsets when it acts as a consumer of a RabbitMQ queue or stream. Valid values are at-least (GPSS stores the offsets before commit), at-most (GPSS stores the offsets after commit), and none. For streams, GPSS also supports strong consistency. The default value is at-least. Refer to Understanding RabbitMQ Message Offset Management for more detailed information.
fallback_offset: earliest | latest: When reading from a RabbitMQ stream, specifies the behaviour of GPSS when it detects a message offset gap. When set to earliest, GPSS automatically resumes a load operation from the earliest available published message. When set to latest, GPSS loads only new messages to the RabbitMQ stream.
save_failing_batch: boolean: Determines whether or not GPSS saves data into a backup table before it writes the data to Greenplum Database. Saving the data in this manner aids recovery when GPSS encounters errors during the evaluation of expressions. The default is false; GPSS does not use a backup table, and returns immediately when it encounters an expression error. When you set this property to true, GPSS writes both the good and the bad data in the batch to a backup table named gpssbackup_<jobhash>, and continues to process incoming data. You must then manually load the good data from the backup table into Greenplum or set recover_failing_batch (Beta) to true to have GPSS automatically reload the good data.

Note
Using a backup table to hedge against mapping errors may impact performance, especially when the data that you are loading has not been cleaned.

recover_failing_batch: boolean (Beta): When set to true and save_failing_batch is also true, GPSS automatically reloads the good data in the batch and retains only the error data in the backup table. The default value is false; GPSS does not process the backup table.

Note
Enabling this property requires that GPSS has the Greenplum Database privileges to create a function.

data_content:: The RabbitMQ message value data type, field names, and type-specific properties. You must specify all RabbitMQ data elements in the order in which they appear in the RabbitMQ message.
column_spec: The source to Greenplum column mapping. The supported column specification differs for different data formats as described below.; The default source-to-target data mapping behaviour of GPSS is to match a column name as defined in source_column_name, column:name, or columns:name with a column name in the target Greenplum Database table. You can override the default mapping by specifying a mapping: block.
data_format: The format of the value data. You may specify a data_format of binary, csv, custom, delimited, or json for the value, with some restrictions.
binary: When you specify the binary data format, GPSS reads the data into a single bytea-type column.
source_column_name: column_name: The name of the single bytea-type column into which GPSS reads the value data.
csv: When you specify the csv data format, GPSS reads the data into the list of columns that you specify. The message content cannot contain line ending characters (CR and LF).
columns:: A set of column name/type mappings. The value [] specifies all columns.
name: column_name: The name of a value column. column_name must match the column name of the target Greenplum Database table.
type: column_data_type: The data type of the column. You must specify an equivalent data type for each data element and the associated Greenplum Database table column.
delimiter: delim_char: Specifies a single ASCII character that separates columns within each message or row of data. The default delimiter is a comma ( ,).
quote: quote_char: Specifies the quotation character. Because GPSS does not provide a default value for this property, you must specify a value.
null_string: nullstr_val: Specifies the string that represents the null value. Because GPSS does not provide a default value for this property, you must specify a value.
escape: escape_char: Specifies the single character that is used for escaping data characters in the content that might otherwise be interpreted as row or column delimiters. Make sure to choose an escape character that is not used anywhere in your actual column data. Because GPSS does not provide a default value for this property, you must specify a value.
force_not_null: columns: Specifies a comma-separated list of column names to process as though each column were quoted and hence not a NULL value. For the default null_string (nothing between two delimiters), missing values are evaluated as zero-length strings.
fill_missing_fields: boolean: Specifies the action of GPSS when it reads a row of data that has missing trailing field values (the row has missing data fields at the end of a line or row). The default value is false, GPSS returns an error when it encounters a row with missing trailing field values.

If set to true, GPSS sets missing trailing field values to NULL. Blank rows, fields with a NOT NULL constraint, and trailing delimiters on a line will still generate an error.

custom

When you specify the custom data format, GPSS uses the custom formatter that you specify to process the input data before writing it to Greenplum Database.

columns:

A set of column name/type mappings. The value [] specifies all columns.

name: column_name

The name of a key or value column. column_name must match the column name of the target Greenplum Database table.

type: column_data_type

The data type of the column. You must specify an equivalent data type for each data element and the associated Greenplum Database table column.

name: formatter_name

When you specify the custom data format, formatter_name is required and must identify the name of the formatter user-defined function that GPSS should use when loading the data.

options:

A set of function argument name=value pairs.

optname=optvalue

The name and value of the set of arguments to pass into the formatter_name UDF.

delimited

When you specify the delimited data format, GPSS reads the data into the list of columns that you specify. You must specify the data delimiter.

columns:

A set of column name/type mappings. The value [] specifies all columns.

name: column_name

The name of a value column. column_name must match the column name of the target Greenplum Database table.

type: column_data_type

The data type of the column. You must specify an equivalent data type for each data element and the associated Greenplum Database table column.

delimiter: delimiter_string

When you specify the delimited data format, delimiter_string is required and must identify the data element delimiter. delimiter_string may be a multi-byte value, and up to 32 bytes in length. It may not contain quote and escape characters.

eol_prefix: prefix_string

Specifies the prefix before the end of line character ( \n) that indicates the end of a row. The default prefix is empty.

quote: quote_char

Specifies the single ASCII quotation character. The default quote character is empty.

If you do not specify a quotation character, GPSS assumes that all columns are unquoted. If you do not specify a quotation character and do specify an escape character, GPSS assumes that all columns are unquoted and escapes the delimiter, end-of-line prefix, and escape itself.

When you specify a quotation character, you must specify an escape character. GPSS reads any content between quote characters as-is, except for escaped characters.

escape: escape_char

Specifies the single ASCII character used to escape special characters (for example, the delimiter, eol_prefix, quote, or escape itself). Therdefault escape character is empty.

When you specify an escape character and do not specify a quotation character, GPSS escapes only the delimiter, end-of-line prefix, and escape itself.

When you specify both an escape character and a quotation character, GPSS escapes only these characters.

json

When you specify the json data format, GPSS can read the data as a single JSON object or as a single JSON record per line.

column:

A single column name/type mapping.

name: column_name

The name of the key or value column. column_name must match the column name of the target Greenplum Database table.

type: json | jsonb | gp_jsonb (Beta) | gp_json (Beta)

The data type of the column.

is_jsonl: boolean

Identifies whether or not GPSS reads the JSON data as a single object or single-record-per-line. The default is false, GPSS reads the JSON data as a single object.

newline: newline_str

A string that specifies the new line character(s) that end each JSON record. The default newline is "\n".

meta:

The data type and field name of the RabbitMQ meta data. meta: must specify the json or jsonb (Greenplum 6 only) data format, and a single json-type column.

The available RabbitMQ meta data properties for a streaming source include:

stream (text) - the RabbitMQ stream name
offset (bigint) - the message offset

The available RabbitMQ meta data properties for a queue source include:

queue (text) - the RabbitMQ queue name
messageId (text) - the message identifier
correlationId (text) - the correlation identifier
timestamp (bitint) - the time that the message was added to the RabbitMQ queue

You can load any of these properties into the target table with a mapping, or use a property in the update or merge criteria for a load operation.

encoding: char_set

The source data encoding. You can specify an encoding character set when the source data is of the csv, custom, delimited, or json format. GPSS supports the character sets identified in Character Set Support in the Greenplum Database documentation.

transformer:

Input data transform block. An input data transformer is a plugin, a set of go functions that transform the data after it is read from the source. The semantics of the transform are function-specific. You specify the library and function names in this block, as well as the properties that GPSS passes to these functions:

path: path_to_plugin_transform_library: The file system location of the plugin transformer library on the Greenplum Streaming Server server host.
on_init: plugin_transform_init_name: The name of an initialization function that GPSS calls when it loads the transform library.
transform: plugin_transform_name: The name of the transform function. GPSS invokes this function for every message it reads.
properties: plugin_transform_property_name: property_value: One or more property name and value pairs that GPSS passes to plugin_transform_init_name and plugin_transform_name.

properties:

RabbitMQ configuration property names and values.

rmq_property_name: The name of a RabbitMQ property.
rmq_property_value: The RabbitMQ property value.

task:

The batch size and commit window.

batch_size:: Controls how GPSS commits data to Greenplum Database. You may specify both max_count and interval_ms as long as both values are not zero ( 0). Try setting and tuning interval_ms to your environment; introduce a max_count setting only if you encounter high memory usage associated with message buffering.
max_count: number_of_rows: The number of rows to batch before triggering an INSERT operation on the Greenplum Database table. The default value of max_count is 0, which instructs GPSS to ignore this commit trigger condition.
interval_ms: wait_time: The minimum amount of time to wait (milliseconds) between each INSERT operation on the table. The default value is 5000.
idle_duration_ms: idle_time: The maximum amount of time to wait (milliseconds) for the first batch of data. When you use this property to enable lazy load, GPSS waits until RabbitMQ data is available before locking the target Greenplum table. You can specify:

0 (lazy load is deactivated)
-1 (lazy load is activated, the job never stops), or
a positive value (lazy load is activated, the job stops after idle_time duration of no data in the RabbitMQ queue or stream)

The default value is 0.

window_size: num_batches: The number of batches to read before running window_statement. The default batch interval is 0.
window_statement: udf_or_sql_to_run: A user-defined function or SQL command(s) that you want to run after GPSS reads window_size number of batches. The default is null, no command to run.
prepare_statement: udf_or_sql_to_run: A user-defined function or SQL command(s) that you want GPSS to run before it executes the job. The default is null, no command to run.
teardown_statement: udf_or_sql_to_run: A user-defined function or SQL command(s) that you want GPSS to run after the job stops. GPSS runs the function or command(s) on job success and job failure. The default is null, no command to run.

option: Properties

schedule:

Controls the frequency and interval of restarting jobs.

retry_interval: retry_time: The period of time that GPSS waits before retrying a failed job. You can specify the time interval in day ( d), hour ( h), minute ( m), second ( s), or millisecond ( ms) integer units; do not mix units. The default retry interval is 5m (5 minutes).
max_retries: num_retries: The maximum number of times that GPSS attempts to retry a failed job. The default is 0, do not retry. If you specify a negative value, GPSS retries the job indefinitely.
running_duration: run_time: The amount of time after which GPSS automatically stops a job. GPSS does not automatically stop a job by default.
auto_stop_restart_interval: restart_time: The amount of time after which GPSS restarts a job that it stopped due to reaching running_duration.
max_restart_times: num_restarts: The maximum number of times that GPSS restarts a job that it stopped due to reaching running_duration. The default is 0, do not restart the job.
quit_at_eof_after: clock_time: The clock time after which GPSS stops a job every day when it encounters an EOF. By default, GPSS does not automatically stop a job that reaches EOF. GPSS never stops a job when the current time is before clock_time, even when GPSS encounters an EOF.

alert:

Controls notification when a job is stopped for any reason (success, completion, error, user-initiated stop).

command: command_to_run: The program that the GPSS server runs on the GPSS server host, including arguments. The command must be executable by GPSS.; command_to_run has access to job-related environment variables that GPSS sets, including: $GPSSJOB_NAME, $GPSSJOB_STATUS, and $GPSSJOB_DETAIL.
workdir: directory: The working directory for command_to_run. The default working directory is the directory from which you started the GPSS server process. If you specify a relative path, it is relative to the directory from which you started the GPSS server process.
timeout: alert_time: The amount of time after a job stops, prompting GPSS to trigger the alert (and run command_to_run). You can specify the time interval in day ( d), hour ( h), minute ( m), or second ( s) integer units; do not mix units. The default alert timeout is -1s (no timeout).

Template Variables

GPSS supports using template variables to specify property values in the load configuration file.

You specify a template variable value in the load configuration file as follows:

<property>: {{<template_var>}}

For example:

max_retries: {{numretries}}

GPSS substitutes the template variable with a value that you specify via the -p | --property <template_var=value> option to the gpsscli dryrun, gpsscli submit, or gpsscli load command.

For example, if the command line specifies:

--property numretries=10

GPSS substitutes occurrences of {{numretries}} in the load configuration file with the value 10 before submitting the job, and uses that value while the job is running.

Notes

If you created a database object name using a double-quoted identifier (delimited identifier), you must specify the delimited name within single quotes in the load configuration file. For example, if you create a table as follows:

CREATE TABLE "MyTable" (c1 text);

Your YAML configuration file would refer to the table name as:

targets:
- gpdb:
    tables:
      - table: '"MyTable"'

You can specify backslash escape sequences in the CSV delimiter, quote, and escape options. GPSS supports the standard backslash escape sequences for backspace, form feed, newline, carriage return, and tab, as well as escape sequences that you specify in hexadecimal format (prefaced with \x). Refer to Backslash Escape Sequences in the PostgreSQL documentation for more information.

Examples

Load data from RabbitMQ as defined in the Version 3 configuration file named loadfromrmq_v3.yaml:

gpsscli load loadfromrmq_v3.yaml

Example loadfromrmq_v3.yaml configuration file:

version: v3
targets:
- gpdb:
    host: mdw-1
    port: 15432
    user: gpadmin
    password: changeme
    database: testdb
    work_schema: public
    error_limit: 25
    tables:
      - table: tbl1
        schema: public
        mode:
          insert {}

sources:
- rabbitmq:
    server: gpadmin:changeme@mdw-1:5672
    queue: test
    vhost: gpadmin
    data_content:
      csv:
        columns: []
        delimiter: ","
        quote: "'"
        null_string: "NA"
        escape: '\'
        force_not_null: "c1,c2"
        fill_missing_fields: true
    properties:
      eof.when.idle: 1500

option:
  schedule: {}

Synopsis

Description

Keywords and Values

Template Variables

Notes

Examples

See Also