Question

考虑以下yaml

hadoop:  
   storage: '/x/y/z/a/b'
   streaming_jar_path: '/x/c/d/f/r/*.jar'
   commands:  
       mkdir: 'hadoop dfs -mkdir dir'
       copyFromLocal: 'hadoop dfs -copyFromLocal from_path to_path'  
       run: 'hadoop jar $streaming_jar_path -mapper mapper_path -reducer reducer_path -input hdfs_input -output hdfs_output'

我想将streaming_jar_path的值替换为$streaming_jar_path，我该怎么做？

我知道我们可以merge hashes使用&(anchors)，但在这里我只想更改一个值
我很抱歉，如果这是微不足道的事情，我对YAML

很新

谢谢

Answer 1

您可以重新构建YAML文件并使用Ansible执行。

commands.yml：

- hosts: localhost
  vars:
    streaming_jar_path: '/x/c/d/f/r/*.jar'
  tasks:
    - name: mkdir
      shell: "hadoop dfs -mkdir dir"
    - name: copyFromLocal
      shell: "hadoop dfs -copyFromLocal from_path to_path"
    - name: run
      shell: "hadoop jar {{ streaming_jar_path }} -mapper mapper_path -reducer reducer_path -input hdfs_input -output hdfs_output"

然后只需运行ansible-playbook即可执行shell命令：

ansible-playbook commands.yml

Answer 2

这应该是一个简单的过程，用于读取文件，编辑数据和写回文件。

import yaml

infile = 'input.yaml'
outfile = 'output.yaml'

#read raw yaml data from file into dict
with open(infile, 'r') as f:
    data = yaml.load(f.read())

#make changes to dict
data['hadoop']['streaming_jar_path'] = '$streaming_jar_path'

#write dict back to yaml file
with open(outfile, 'w') as f:
    f.write(yaml.dump(data))

Answer 3

这是我的解决方法，使用python的字符串格式化程序......

yaml档案......

hadoop:  
   streaming_jar_path: '/x/c/d/f/r/*.jar'
   commands:  
       run: 'hadoop jar {} -mapper mapper_path -reducer reducer_path -input hdfs_input -output hdfs_output'

python代码：

s_jar_p = params['hadoop']['streaming_jar_path'] 
run_cmd = (params['hadoop']['commands']['run']).format(s_j_p)

这对我有用： - ）

YAML - 用变量代替值？

3 个答案: