Getting multi-line after converting json to yaml

[msgongora]

Hi folks,

I spent a lot of time searching and trying out multiple options without success. I hope someone has dealt with this before and can help me.
I'm converting an API Response (JSON) to YAML and in the process, I would need to filter out some fields leaving only some of them in my final YAML. One of the fields (.view.query) in the final result is a string that for better management, it needs to be in a multi-line format.

cat view.json
{
  "kind": "bigquery#table",
  "etag": "U2ZBhyVSbRj0w==",
  "id": "projecta:datasetbar.sampleview",
  "numBytes": "0",
  "numLongTermBytes": "0",
  "numRows": "0",
  "creationTime": "1606239345662",
  "lastModifiedTime": "1680742293586",
  "type": "VIEW",
  "view": {
    "query": "(SELECT\n  Month,\n  Date,\n  I,\n  CT,\n  C,\n  D,\n  OS,\n\n  (CASE\n    WHEN lower(C) like '%discovery%' then TC + VT \n    Else TC \n    End) as UTC \nFROM\n  `projecta.datasetbar.tablefoo`\nWHERE\n  date>='2020-10-01'\n  AND date<= '2020-12-31' \n  Order by 2,3 )\n  \n  \n  \nUnion all\n\n(SELECT *  FROM `projecta.datasetbar.tablezoo` where C>=0)",
    "useLegacySql": false
  },
  "numLongTermPhysicalBytes": "0"
}

yq e '.|= pick(["friendlyName", "description", "labels", "view"])' view.json -oy

view:
  query: "(SELECT\n  Month,\n  Date,\n  I,\n  CT,\n  C,\n  D,\n  OS,\n\n  (CASE\n    WHEN lower(C) like '%discovery%' then TC + VT \n    Else TC \n    End) as UTC \nFROM\n  `projecta.datasetbar.tablefoo`\nWHERE\n  date>='2020-10-01'\n  AND date<= '2020-12-31' \n  Order by 2,3 )\n  \n  \n  \nUnion all\n\n(SELECT *  FROM `projecta.datasetbar.tablezoo` where C>=0)"
  useLegacySql: false

This is what I'd like to get:

view
  query: |-
    (SELECT
      Month,
      Date,
      I,
      CT,
      C,
      D,
      OS,

      (CASE
        WHEN lower(C) like '%discovery%' then TC + VT
        Else TC
        End) as UTC
    FROM
      `projecta.datasetbar.tablefoo`
    WHERE
      date>='2020-10-01'
      AND date<= '2020-12-31'
      Order by 2,3 )



    Union all

    (SELECT *  FROM `projecta.datasetbar.tablezoo` where C>=0)

I realized yq returns multi-line when I'm getting only the query value

yq e '.view.query' view.json -oy
(SELECT
  Month,
  Date,
  I,
  CT,
  C,
  D,
  OS,

  (CASE
    WHEN lower(C) like '%discovery%' then TC + VT
    Else TC
    End) as UTC
FROM
  `projecta.datasetbar.tablefoo`
WHERE
  date>='2020-10-01'
  AND date<= '2020-12-31'
  Order by 2,3 )



Union all

(SELECT *  FROM `projecta.datasetbar.tablezoo` where C>=0)

but I can't figure out how to get the expected result from a single yq execution. I would appreciate it if someone has some input here.

[mikefarah]

This caused by an annoying existing issue in the underlying go-yaml parser whereby if there are any trailing spaces before the newline character, then the entire string is printed in double quotes (see go-yaml/yaml#880)

Luckily, you can use the sub operator to fix this on the fly:

yq '.view.query |= sub(" *\n", "\n")' file.json -oy

[msgongora]

Thank you for taking the time to answer @mikefarah. This was definitely the missing part.

Here is what I ended up doing in my use case.

yq '.view.query |= sub(" *\n", "\n") |
    pick(["friendlyName", "description", "labels", "view"])'  view.json -oy