如何在AWS状态语言（ASL）中复用JSON代码？

我有一个包含14个状态的Step Function工作流，每个状态除命令外完全一致。能否在ASL中复用该状态，仅修改命令内容，避免生成包含大量重复条目的冗长JSON文件？

原工作流代码示例（简化并修正语法后）：

{
  "Comment": "ECS RunTask Jobs",
  "StartAt": "ECS RunTask 01",
  "States": {
    "ECS RunTask 01": {
      "Type": "Task",
      "Resource": "{some resource}",
      "Parameters": {"LaunchType": "FARGATE", ...},
      "Next": "ECS RunTask 02",
      "ResultSelector": { "commands": ["command1"] },
      "Catch": [{
        "ErrorEquals": ["States.ALL"],
        "Next": "Send Email Error"
      }]
    },
    "ECS RunTask 02": {
      "Type": "Task",
      "Resource": "{some resource}",
      "Parameters": {"LaunchType": "FARGATE", ...},
      "Next": "ECS RunTask 03",
      "ResultSelector": { "commands": ["command2"] },
      "Catch": [{
        "ErrorEquals": ["States.ALL"],
        "Next": "Send Email Error"
      }]
    },
    // ... 重复到ECS RunTask 14
    "Send Email Error": {
      "Type": "Task",
      "Resource": "...",
      "End": true
    },
    "Send Email Complete": {
      "Type": "Task",
      "Resource": "...",
      "End": true
    }
  }
}

当然可以，ASL提供两种核心方式来复用状态，彻底减少重复代码：

方法1：状态引用（State References） #

通过$Ref语法定义一个基础状态模板，然后在每个任务实例中引用模板并仅覆盖差异字段（比如命令和跳转目标）。

示例代码：

{
  "Comment": "ECS RunTask Jobs with State References",
  "StartAt": "ECS RunTask 01",
  "States": {
    // 定义可复用的基础模板（用下划线标记为模板，不会被直接执行）
    "_ECS RunTask Template": {
      "Type": "Task",
      "Resource": "{some resource}",
      "Parameters": {"LaunchType": "FARGATE", ...},
      "Catch": [{
        "ErrorEquals": ["States.ALL"],
        "Next": "Send Email Error"
      }]
    },
    // 引用模板并覆盖差异字段
    "ECS RunTask 01": {
      "$Ref": "#/States/_ECS RunTask Template",
      "Next": "ECS RunTask 02",
      "ResultSelector": { "commands": ["command1"] }
    },
    "ECS RunTask 02": {
      "$Ref": "#/States/_ECS RunTask Template",
      "Next": "ECS RunTask 03",
      "ResultSelector": { "commands": ["command2"] }
    },
    // ... 依次定义到ECS RunTask 14
    "ECS RunTask 14": {
      "$Ref": "#/States/_ECS RunTask Template",
      "Next": "Send Email Complete",
      "ResultSelector": { "commands": ["command14"] }
    },
    "Send Email Error": {
      "Type": "Task",
      "Resource": "...",
      "End": true
    },
    "Send Email Complete": {
      "Type": "Task",
      "Resource": "...",
      "End": true
    }
  }
}

这种方式适合需要严格按固定顺序执行任务的场景，每个实例仅需指定跳转目标和命令，其余逻辑完全复用模板。

方法2：Map状态批量执行 #

如果14个命令是独立可批量处理的任务，推荐用Map状态彻底消除重复状态定义。通过遍历命令列表，自动执行同一个任务逻辑。

示例代码：

{
  "Comment": "ECS RunTask Jobs with Map State",
  "StartAt": "Run All ECS Tasks",
  "States": {
    "Run All ECS Tasks": {
      "Type": "Map",
      "ItemsPath": "$.commands", // 指向输入中的命令列表
      "MaxConcurrency": 1, // 设为1则顺序执行，大于1则并行执行
      "Parameters": {
        "command.$": "$$.Map.Item.Value"
      },
      "Iterator": {
        "StartAt": "ECS RunTask",
        "States": {
          "ECS RunTask": {
            "Type": "Task",
            "Resource": "{some resource}",
            "Parameters": {
              "LaunchType": "FARGATE",
              "Command.$": "$.command" // 动态引用当前遍历到的命令
            },
            "Catch": [{
              "ErrorEquals": ["States.ALL"],
              "Next": "Send Email Error"
            }],
            "End": true
          }
        }
      },
      "Next": "Send Email Complete",
      "Catch": [{
        "ErrorEquals": ["States.ALL"],
        "Next": "Send Email Error"
      }]
    },
    "Send Email Error": {
      "Type": "Task",
      "Resource": "...",
      "End": true
    },
    "Send Email Complete": {
      "Type": "Task",
      "Resource": "...",
      "End": true
    }
  }
}

工作流输入示例（包含所有命令）：

{
  "commands": ["command1", "command2", ..., "command14"]
}

这种方式的优势是代码极度简洁，新增或删除命令只需修改输入列表，无需调整状态结构。

内容的提问来源于stack exchange，提问作者John Garner