Variables-Data Management-阿里云帮助中心

Variable overview

Variables are dynamic values in a task flow, categorized as system variables, time variables, and non-time variables. You can reference any variable within a node using the${varName} syntax.

System variables

System variables are predefined values, such as the current task flow ID and the run time of the task flow.

Note

If a custom variable shares a name with a system variable, its value overrides the system variable's value, giving the custom variable higher priority.

Time variables

Task flow variable (global variable): A task flow variable is calculated by applying a time offset to the data timestamp. By default, the data timestamp is the run time minus one day. As a global variable, it is available to all nodes in the current task flow.

Note
You can configure a task flow variable on any node in a task flow. The configuration is automatically synchronized to all other nodes.
Node variable: A time variable created by applying a time offset to the built-in bizdate (T-1) variable. A node variable is scoped to its defining node.

For example, if the current date is August 22, 2023, and you apply a time offset of +1 day, the variable's value becomes August 23, 2023.

Non-time variables

Task flow constant (global variable): A task flow constant is a fixed value that can be used in all nodes within the current task flow.
Input variable: The system automatically populates input variables. You can reference an input variable in the SQL statements of the current node by using the ${var_name} syntax or as a conditional variable in a Conditional Branch node. Input variables include:
- Upstream variable: An output variable from an upstream node.
- Status variable: For more information, see Status variables.
- System variable: For more information, see System variables.
Output variable: You define and assign a value to an output variable on the current node. Downstream nodes can then access and reference it.

For example, if you configure an output variable in a Script node, it can be referenced in the SQL statements of a downstream node.
Output variable from a Single Instance SQL node: You can use a Single Instance SQL node to define a result set from an SQL query. This result set forms a two-dimensional matrix. During variable definition, you can extract a single element or concatenate all values from a row or column.
Script output variable: If the last line of a script's output is a JSON object in the format{ key1: value1, key2: value2, … }, and the values are of the String type, the script task automatically parses the JSON key-value pairs. This creates a variable namedkey, and the value of the variablekey is set to the corresponding value. You can reference the variable as${key}.

For example, if the last line of the script isecho {"hello": "world"}, the script task creates a variable namedhello, and the value ofhello is world.

Configure time variables

Parameter

Description

Variable name

Enter the name of the custom variable.

Note

To delete a configured variable, click the 删除-变量 icon.

Variable rule

Configure the rules for the time variable.

Time format: Specify the desired time format. For more information, see Time formats.
Time offset: A time variable is calculated by applying an offset to the data timestamp (bizdate), which defaults to the day before the task flow's run time.

For example, consider a variable named6_month_ago with the time format yyyy-MM-dd and an offset of -6 months. If the run time is 2021-08-12 (making bizdate 2021-08-11), the value of the${6_month_ago} variable is 2021-02-11.

Using month or year as the offset unit can result in non-continuous date sequences. For a continuous daily sequence, use day as the offset unit.

Note

After you configure a time variable, you can reference it in SQL statements using the${var_name} syntax. You can also click SQL Preview to check the time variable's value.

Time formats

The following table describes the supported time format specifiers.

Specifier	Description	Example input	Example output
Era	`G` indicates the era.	Gyyyy	AD 2021
Year	`y` or `yyyy`: The year. `yy`: The last two digits of the year. `Y`: The week-numbering year, which is the year at the end of the week (Sunday).	yyyy	2021
Month	`M`: The month of the year. Valid values: [1,12]. `MM`: Zero-padded. Valid values: [01,12]. `MMM` returns the month name, such as "January" to "December".	MM	08
Week	`w`: The week of the year. Valid values: [1, 52]. `ww`: Zero-padded. Valid values: [01, 52]. `W`: The week of the month. Valid values: [1, 5].	ww	13
Day	`D`: The day of the year. Valid values: [1, 365]. `DD`: Zero-padded. Valid values: [01, 365]. `DDD`: Padded with two leading zeros. Valid values: [001, 365]. `d`: The day of the month. Valid values: [1, 31]. `dd`: Zero-padded. Valid values: [01, 31].	D	360
Day of week	`E`: The name of the day of the week, from Monday to Sunday. `e`: The number of the day of the week. Valid values: [1, 7]. By default, 1 represents Monday. Note You can set the time offset to +1 day to align with calendars where 1 represents Sunday.	e	1
AM/PM	`a`: The AM/PM marker. Returns AM for times from 00:00 to 11:59, and PM for times from 12:00 to 23:59.	a	AM
Hour	`H`: The hour of the day (0-23). `0` represents midnight. Valid values: [0, 23]. `HH`: Zero-padded. Valid values: [00, 23]. `h`: The hour in AM/PM (1-12). `1` represents the first hour. Valid values: [1, 12]. `hh`: Zero-padded. Valid values: [01, 12]. `K`: The hour in AM/PM (0-11). `0` represents the first hour. Valid values: [0, 11]. `KK`: Zero-padded. Valid values: [00, 11]. `k`: The hour of the day (1-24). `1` represents the first hour. Valid values: [1, 24]. `kk`: Zero-padded. Valid values: [01, 24].	HH	10
Minute	`m`: The minute of the hour. Valid values: [0, 59]. `mm`: Zero-padded. Valid values: [00, 59].	m	27
Second	`s`: The second of the minute. `S`: The millisecond.	ss	08
Time zone	`z`: The time zone.	z	UTC+08:00

Examples of combined time formats:

Example input	Example output
`yyyy-MM-dd`	2021-08-12
`yyyyMM01`	20210801
`HH:mm:ss`	11:05:21
`yyyyMMdd HH:mm:ss`	20210812 11:05:21

System variables

Variable name	Description	Example
sys.flow.start.timestamp	The run time's timestamp.	2021-05-24T11:20:07.562+08:00
sys.flow.start.year	The run time's year.	2021
sys.flow.start.month	The run time's month.	5
sys.flow.start.day	The run time's day.	24
sys.flow.start.hour	The run time's hour.	11
sys.flow.start.minute	The run time's minute.	20
sys.flow.start.second	The run time's second.	7
sys.flow.start.milliseconds	The run time's milliseconds.	562
sys.flow.start.timezone	The run time's time zone.	Asia/Shanghai
sys.flow.biztime	The data timestamp. By default, this is the run time minus one day.	1621740007562
sys.flow.name	The task flow name.	dwd_activity_daily_pv
sys.node.name	The node name.	Single Instance SQL-1

Status variables

Variable name	Description
all_success	All upstream tasks ran successfully.
all_failed	All upstream tasks failed.
one_success	At least one upstream task ran successfully.
one_failed	At least one upstream task failed.

Note

You can use status variables in a Conditional Branch node to control whether the task flow runs subsequent tasks based on a specific run status.