Using Dasel to export JSON at different levels to CSV

[eroth]

Hi, thanks so much for making Dasel! I'm trying to use it to export JSON at different levels (heterogenous levels?) to CSV and am wondering if it can do what I'm looking for. I've looked a bunch here and in various blog posts but can't seem to find anything or figure it out. This is a sample of the structure I'm working with (an array of these objects pulled from GH's GraphQL API):

[
{
  "author" : {
    "login" : "querty"
  },
  "number" : 4521,
  "createdAt" : "2022-11-21T02:57:35Z",
  "mergedAt" : "2022-12-05T02:16:18Z",
  "mergedBy" : {
    "login" : "querty"
  },
  "approvers" : {
    "nodes" : [
      {
        "author" : {
          "name" : "Bill Blass",
          "login" : "bblass"
        },
        "state" : "APPROVED"
      },
      {
        "author" : {
          "name" : "Ricky Eckhart",
          "login" : "wreck-hart"
        },
        "state" : "APPROVED"
      }
    ]
  },
  "title" : "Does something really amazing and magical",
  "repository" : {
    "owner" : {
      "login" : "orgo"
    },
    "name" : "my-my-my"
  }
},
{
...
}
]

What I'm trying to do is only pull out certain fields at different levels, then export them to CSV. What I'd like to end up with is something like the below, where author comes from author.login and everything else is at the same level of the structure:

author,number,title,createdAt,mergedAt
"querty",4521,"Does something really amazing and magical","2022-11-21T02:57:35Z","2022-12-05T02:16:18Z"
...

I'm able to print out one value, e.g., all values for author, but can't seem to figure out how to map a custom structure out of this and export it to CSV like what I'm after. Any help would be appreciated!

[TomWright]

This sounds like a good use-case for mapOf.

$ dasel -f file.json -w csv 'all().mapOf(author,author.login,number,number,title,title).merge()'
author,number,title
querty,4521.000000,Does something really amazing and magical

[TomWright]

Hey @chmac, you can find the mapOf docs here: https://daseldocs.tomwright.me/functions/mapof

In short though, the arguments are in pairs of key and value.

In the example above the map key is author and the value will be whatever is resolved from the author.login selector.

Good shout RE the docs - I'll see if I can get the mapOf docs improved today

[TomWright]

I've updated the docs: https://daseldocs.tomwright.me/functions/mapof
Please let me know if you think there could be any other improvements

[chmac]

@TomWright The updated docs look awesome, I'm impressed by the near instant response! :-)

In the example about outputting to CSV, I'd suggest outputting both name and phone number, so that it's clearer to people how to get 2 fields in the CSV. Now that I see your examples I realise that the first of the 2 pairs in mapOf is a string and doesn't need to exist in the data, that makes sense!

[chmac]

What I had in mind was replacing this:

dasel -r json -w csv 'all().mapOf(phone,phone).merge()'

With something like this:

dasel -r json -w csv 'all().mapOf(firstName,name.first,phone,phone).merge()'

Then hopefully the output would be:

firstName,phone
Tom,07
Joe,07

[TomWright]

Good idea @chmac , I've just made that change