Read YAML file from Bash script

parse_yaml.sh

#!/bin/sh
parse_yaml() {
   local prefix=$2
   local s='[[:space:]]*' w='[a-zA-Z0-9_]*' fs=$(echo @|tr @ '\034')
   sed -ne "s|^\($s\)\($w\)$s:$s\"\(.*\)\"$s\$|\1$fs\2$fs\3|p" \
        -e "s|^\($s\)\($w\)$s:$s\(.*\)$s\$|\1$fs\2$fs\3|p"  $1 |
   awk -F$fs '{
      indent = length($1)/2;
      vname[indent] = $2;
      for (i in vname) {if (i > indent) {delete vname[i]}}
      if (length($3) > 0) {
         vn=""; for (i=0; i<indent; i++) {vn=(vn)(vname[i])("_")}
         printf("%s%s%s=\"%s\"\n", "'$prefix'",vn, $2, $3);
      }
   }'
}

test.sh

#!/bin/sh

# include parse_yaml function
. parse_yaml.sh

# read yaml file
eval $(parse_yaml zconfig.yml "config_")

# access yaml content
echo $config_development_database

zconfig.yml

development:
  adapter: mysql2
  encoding: utf8
  database: my_database
  username: root
  password:

This is fantastic! Thank you all who have contributed to this. It was exactly what I was looking for. I think it was mentioned above by @wadewegner, it has issues with hyphenated hash keys such as proj-ect:. wadewegner you said that your first pure bash change worked for you, but had array issues. I was unsuccessful with using your approach. And unfortunately I am not able to introduce other tooling like python packages such as niet into my project. Has anyone been able to get this working for hyphenated keys?

How could i "echo" all given variables?

Thank you very much!

Trying to use Martins modified parser from stckovrflow. Input file:

schemas:
- name: exports
  tables:
  - name: wolverine
    description: sample message here
      may overflow next line
    active_date: 2019-01-07 00:00:00
    columns:
    - name: height
      type: timestamp without time zone
    - name: strength
      type: bigint
      description: sample message here
        may overflow next line
      example: 21352352
    - name: power
      type: bigint
      description: sample message here
        may overflow next line
      example: 10001
  - name: cyclops
    description: sample message here
      may overflow next line
    active_date: 2018-12-15 00:00:00
    columns:
    - name: size
      type: datetime
      description: sample message here
        may overflow next line
      example: 2018-03-03 12:30:00
    - name: power
      type: timestamp without time zone
      description: sample message here
        may overflow next line

Parse function:

function parse_yaml {
   local prefix=$2
   local s='[[:space:]]*' w='[a-zA-Z0-9_]*' fs=$(echo @|tr @ '\034')
   sed -ne "s|,$s\]$s\$|]|" \
        -e ":1;s|^\($s\)\($w\)$s:$s\[$s\(.*\)$s,$s\(.*\)$s\]|\1\2: [\3]\n\1  - \4|;t1" \
        -e "s|^\($s\)\($w\)$s:$s\[$s\(.*\)$s\]|\1\2:\n\1  - \3|;p" $1 | \
   sed -ne "s|,$s}$s\$|}|" \
        -e ":1;s|^\($s\)-$s{$s\(.*\)$s,$s\($w\)$s:$s\(.*\)$s}|\1- {\2}\n\1  \3: \4|;t1" \
        -e    "s|^\($s\)-$s{$s\(.*\)$s}|\1-\n\1  \2|;p" | \
   sed -ne "s|^\($s\):|\1|" \
        -e "s|^\($s\)-$s[\"']\(.*\)[\"']$s\$|\1$fs$fs\2|p" \
        -e "s|^\($s\)-$s\(.*\)$s\$|\1$fs$fs\2|p" \
        -e "s|^\($s\)\($w\)$s:$s[\"']\(.*\)[\"']$s\$|\1$fs\2$fs\3|p" \
        -e "s|^\($s\)\($w\)$s:$s\(.*\)$s\$|\1$fs\2$fs\3|p" | \
   awk -F$fs '{
      indent = length($1)/2;
      vname[indent] = $2;
      for (i in vname) {if (i > indent) {delete vname[i]; idx[i]=0}}
      if(length($2)== 0){  vname[indent]= ++idx[indent] };
      if (length($3) > 0) {
         vn=""; for (i=0; i<indent; i++) { vn=(vn)(vname[i])("_")}
         printf("%s%s%s=\"%s\"\n", "'$prefix'",vn, vname[indent], $3);
      }
   }'
}

Current Output :

1="name: exports"
1_1="name: wolverine"
1_1_description="sample message here"
1_1_active_date="2019-01-07 00:00:00"
1_1_1="name: height"
1_1_1_type="timestamp without time zone"
1_1_2="name: strength"
1_1_2_type="bigint"
1_1_2_description="sample message here"
1_1_2_example="21352352"
1_1_3="name: power"
1_1_3_type="bigint"
1_1_3_description="sample message here"
1_1_3_example="10001"
1_2="name: cyclops"
1_2_description="sample message here"
1_2_active_date="2018-12-15 00:00:00"
1_2_1="name: size"
1_2_1_type="datetime"
1_2_1_description="sample message here"
1_2_1_example="2018-03-03 12:30:00"
1_2_2="name: power"
1_2_2_type="timestamp without time zone"
1_2_2_description="sample message here"

But I need to get the values in nested format csv instead of synthetic numbered markers. The idea is to populate a nested relationship from parent to child elements . Anything in the same level will be written delimited by comma like below:

Wolverine.height,"timestamp without time zone",,
Wolverine.strength,bigint,"sample message here may overflow next line",21352352
Wolverine.power,bigint,"sample message here may overflow next line",10001
Cyclops.size,datetime,"sample message here may overflow next line","2018-03-03 12:30:00"
Cyclops.power,"timestamp without time zone",something,"sample message here may overflow next line",,

How can I reformat it? Can someone help ?

Hi,
I tried to use this in Jenkins for job config parsing.
I ran successfully with Python 2.6.6 but not with Python 2.7.5
Help is really appreciated.

Ran with Python 2.6.6

• echo 'Get Job Properties'
  Get Job Properties
• /usr/bin/python -V
  Python 2.6.6
• cd /bms/webapps/jenkins/workspace/pro-group/test/yaml_parser
• . parse_yaml.sh
  ++ parse_yaml access_info.yaml accessconfig_

Ran with Python 2.7.5

• echo 'Get Job Properties'
  Get Job Properties
• /usr/bin/python -V
  Python 2.7.5
• cd /bms/webapps/jenkins/workspace/pro-group/test/yaml_parser
• . parse_yaml.sh
  /bms/webapps/jenkins/jenkins881879139786658308.sh: line 29: .: parse_yaml.sh: file not found
  Build step 'Execute shell' marked build as failure

Thank you for publishing! Helped save alot of time 👍

Hello,

niet now support eval output format so you can do things like that:

$ pip install -U niet
$ echo '{"foo-biz": "bar", "fizz": {"buzz": ["zero", "one", "two", "three"]}}' | niet -f eval .
 foo_biz="bar";fizz__buzz=( zero one two three )
$ eval $(echo '{"foo-biz": "bar", "fizz": {"buzz": ["zero", "one", "two", "three"]}}' | niet -f eval .)
$ echo ${foo_biz}
bar
$ echo ${fizz__buzz}
zero one two three
$ eval $(echo '{"foo-biz": "bar", "fizz": {"buzz": ["zero", "one", "two", "three"]}}' | niet -f eval '"foo-biz"'); echo ${foo_biz}
bar
$ echo '{"foo-biz": "bar", "fizz": {"buzz": ["zero", "one", "two", "three"]}}' | niet -f eval fizz.buzz
fizz_buzz=( zero one two three );
$ eval $(echo '{"foo-biz": "bar", "fizz": {"buzz": ["zero", "one", "two", "three"]}}' | niet -f eval fizz.buzz)
$ for el in ${fizz_buzz}; do echo $el; done
zero
one
two
three

niet work with JSON and YAML input, also you can convert each format to each other, just by using pip install -U niet.

Thank you. Can we validate that the username as below is mandatory and have value, non empty field. Right now it passed.
development:
adapter: mysql2
encoding: utf8
database: my_database
username:
if anyone tried, please share. Thanks much

https://github.com/jasperes/bash-yaml for impatient !

@pkuczynski, It's work for me as well to read the key in yaml file. Thanks for that. Do you have the script to replace the value after reading? i have to replace the each value which i reads .

Hi! What is the expected output from running test.sh? Perhaps listing an output file above would be useful.

Works great.

nice script!

Works really well, thank you!

description: sample message here
may overflow next line

I have a similar issue how could we store description variable that may overflow into the next line in the same

1_2_1_description="sample message here"
TO
1_2_1_description="sample message here may overflow next line"

Thank you!

I think python is superior in this scripting situation. why on earth anyone would use a bash script to do such a thing? python is generally pre-installed in mac and Linux.

python is so much clean and readable.

pip install pyyaml

---

import yaml

FILENAME = 'your_file.yml'

with open(FILENAME) as file:
    data:dict = yaml.full_load(file)

# data is a python dictionary

The traversal logic is super. Love it !! Sometimes its just fun to do these crazy stuff with Bash.

jq for JSON and yq for YAML
https://github.com/mikefarah/yq

Hi all! This is very interested thread. Anybody knows how to make it working where string spans several lines (through "|" pipe), example:

key1: |
   1st line
   2nd line
   3rd line
key2: |
   1st line
   2nd line
   3rd line

thanks for this.
added this to end after awk as a hack not to interpret $ interpolation (application: reading serverless.yml config into bash)

 | tr "$" "#" # don't want to interpret $

i cant figure out why its not running succesfully for me whenever i execute ./test.sh it gives me error
./test.sh
./test.sh: line 4: .: parse_yaml.sh: file not found
./test.sh: line 7: parse_yaml: command not found

has anybody else faced this issue please help

@Nigam8972 try with
. ./parse_yaml.sh
if they are in the same folder (it's what I use)

Cool, just discovered this; of course the more recent suggestions are more robust and feature-rich, but I found myself needing this on constrained (i.e. busybox, sh and no python) systems where going POSIX-sh only is the way to go.

The method here by @wadewegner (or similar) to pre-process problematic keys (but not their values) was the way to go, i.e. by using "advanced" sed hold-space etc.