We have updated our Terms of Service, Code of Conduct, and Addendum.

Query on simplifying JSON to CSV conversion.

Options

If I am reading the docs correctly, when converting from json to CSV, I have to define all the fields. Since all of the fields are really already known, is there a way to output all the fields first without having to define all of them? We are working on converting some Azure logs, and there are LOTs of properties.field(s). Hoping there is another way without us having to define all of these in the "Fields to serialize" field within the Serialize function.

Answers

  • Jon Rust
    Jon Rust Posts: 439 mod
    Options

    All fields must be listed, in the order you require. Copy-pasta works for the list. Possibly in your analysis tool of choice you can get a list?

  • hgoolya
    hgoolya Posts: 27 mod
    Options

    ok, tahnks

  • hgoolya
    hgoolya Posts: 27 mod
    Options

    thanks

  • Jon Rust
    Jon Rust Posts: 439 mod
    Options

    create a temp function to serialize into K=V and another to use regex to pull all the field names. In the preview window, copy-pasta the list

  • Jon Rust
    Jon Rust Posts: 439 mod
    Options

    hold on, trying it out

  • Jon Rust
    Jon Rust Posts: 439 mod
    Options

    ```{ "conf": { "output": "default", "streamtags": [], "groups": {}, "asyncFuncTimeout": 1000, "functions": [ { "filter": "true", "conf": { "mode": "extract", "type": "json", "srcField": "raw", "dstField": "_raw" }, "id": "serde" }, { "filter": "true", "conf": { "type": "kvp", "fields": [ "!", "!cribl_breaker", "" ], "dstField": "raw", "cleanFields": false, "srcField": "_raw" }, "id": "serialize" }, { "filter": "true", "conf": { "source": "_raw", "iterations": 100, "overwrite": false, "regex": "/(?<fields>[-a-zA-Z0-9]+)=/g" }, "id": "regex_extract" }, { "filter": "true", "conf": { "add": [ { "disabled": false, "name": "fields", "value": "fields.join(',')" } ] }, "id": "eval" } ] }, "id": "list_fields" }```

  • Jon Rust
    Jon Rust Posts: 439 mod
    Options

    » my sample data started as stringified JSON » i parsed it into actual JSON » then serialized into K=V » then regex extract for all field names (strings leading up to `=`), with global flag set » then eval to join the fields array into a list

  • Jon Rust
    Jon Rust Posts: 439 mod
    Options

    obviously nested json would present a problem, but then i don't know how you'd plan on making that CSV in the first place

  • hgoolya
    hgoolya Posts: 27 mod
    Options

    great. Let me give that a try

  • hgoolya
    hgoolya Posts: 27 mod
    Options

    the nested is going to be the issue

  • Jon Rust
    Jon Rust Posts: 439 mod
    Options

    how would you do nested in CSV?

  • hgoolya
    hgoolya Posts: 27 mod
    Options

    not sure. I think we are doing this for another source, so I need to look. We will tackle this next week