This question has also been answered here:
Stackoverflow: Unexpected ConvertTo-Json results? Answer: it has a default -Depth of 2
GitHub: ConvertFrom-Json and ConvertTo-Json breaks arrays
Mircrosoft Docs: ConvertTo-JSON
TL;DR
If you save your .json with ConvertTo-JSON and it break it, you may want to speccify the -Depth parameter, as it's default value it 2 and it wont parse nestet objects beyond that.
Original Post:
So I want to load, then save data from a JSON file through a PowerShell skript. The JSON contains an array within an object, which looks like this:
{
"head": {
"head2": {
"data0": "a",
"data1": "b"
},
"head3": {
"data8": [
"x",
"y",
"z"
],
"data9": "hello"
}
}
}
Notice the array "data8".
Now when I load and save the file like so: Get-Content test.json | ConvertFrom-JSON | ConvertTo-JSON I want to end up with the exact same file, since I'm not changing anything. However the result is this:
{
"head": {
"head2": {
"data0": "a",
"data1": "b"
},
"head3": {
"data8": "x y z",
"data9": "hello"
}
}
}
The "data8" has been reduced to a single string and I struggle to find out why. It seems to happen during the ConvertTo-JSON because when I don't do that yet, It will give me an array of strings containing x,y and z.
E.g. (Get-Content test.json | ConvertFrom-JSON).head.head3.data8 will result in
x
y
z
and (Get-Content test.json | ConvertFrom-JSON).head.head3.data8 -is [array] gives True
Things I've tried:(get-content test.json | convertfrom-json).head.head3.data8
Various encoding methods in test.json, but this even happens when I save the file as utf8 and add -encoding utf8 to Get-Content so I don't believe that is has to do with that.
For the sake of trying, I even added -compress to ConvertTo-JSON which didn't help either.
Okay so apparently PowerShell only parses data up to a depth of 2 by default. So as pointed out Here you have to specify the depth with the depth prameter of ConvertTo-Json