Equivalent of yq --collect --tojson?

[ghostsquad]

Given a yaml file that has multiple documents, e.g.

foo: bar
---
baz: biz

I'd like to get an array of each document. Is this possible with dasel?

[TomWright]

This isn't possible right now but is a fairly simple use-case.

I'll add a --merge-input-documents flag that can do this. E.g.

echo 'foo: bar
---
baz: biz' | dasel -r yaml -w json --merge-input-documents .
[
  {
    "foo": "bar"
  },
  {
    "baz": "biz"
  }
]

The work is done on the merge-documents branch which you can install from with the following:

go install github.com/tomwright/dasel/cmd/dasel@merge-documents

The documentation is here: https://app.gitbook.com/@tomwright/s/dasel/usage/flags/merge-input-documents

[link4531]

Am I correct in assuming this feature could also merge two separate files (documents?) with similar structure?

file1: foo: bar
file2: baz: biz
this is a terrible example
cat file1 file2 | dasel -yaml -json --merge-input-documents

Sorry if I am asking the same question as mentioned above.

Also any chance we could get a Release built with this feature included maybe tagged as 'preview' or 'beta'.
I do not have the ability to utilize the 'go install ...' per your instruction.

[TomWright]

Assuming your documents are correctly formatted and that the cat output is properly formed, yes.

x.yaml

foo: bar
---

y.yaml

baz: biz

Without flag:

$ cat x.yaml y.yaml| ~/go/bin/dasel -p yaml .         
foo: bar
---
baz: biz

With flag:

$ cat x.yaml y.yaml| ~/go/bin/dasel --merge-input-documents -p yaml .
- foo: bar
- baz: biz

This flag tells dasel to merge all the given documents into an array/list, where .[0] contains the first document, .[1] contains the second document and so on.

---

I will merge and release this feature once I know it's named appropriately and doesn't have any bugs (more tests).

[link4531]

Super stoked for this feature.

Question is there a priority to the merge order?
such as one file contains foo: bar and a later file in the array contains foo: biz << I assume the last value to be read will be written in the output?

[TomWright]

That's great to hear! I'm busy at work at the moment but will set aside some time to get this done soon 👍

Dasel just reads and merges the document in the order they are provided, so yes your assumption is correct.

Do you have some basic examples of what you may try to do with this feature? They will give me some good use-cases to test against.

[TomWright]

This has been released under v1.15.0.

[TomWright]

@link4531

Take note that merge in this case means taking all of the input documents and adding them to a single array of those documents.

Input: a, b, c
Output: [a, b, c]

Where:
a: {"number": 1}
b: {"number": 2}
c: {"number": 3}

Output:

[
    {"number": 1},
    {"number": 2},
    {"number": 3}
]

The use of "merge" here could be mistaken in that you may expect the following output:

{
    "number": 3
}

I'm considering adding a new feature that will provide the second output but am not yet sure what the usage will look like.

#133