From Friday, April 19th (11:00 PM CDT) through Saturday, April 20th (2:00 PM CDT), 2024, ni.com will undergo system upgrades that may result in temporary service interruption.
We appreciate your patience as we improve our online experience.
LabVIEW
Using LabVIEW to compose JSON strings incrementally
Solved!
I'm struggling to figure out a way to build JSON representations of objects incrementally in LabVIEW. By this I mean I want to split the generation of a complete JSON document across separate VIs in order to keep my code DRY.
In particular, I'm creating an application in LabVIEW that needs to communicate with a webservice. This webservice expects to receive JSON patch operations (see RFC 6902), which means JSON objects with the following schema:
{
"op": <enumeration: add, replace, etc.>,
"path": <string>,
"value": <arbitrarily complex object>
}
It makes sense to create one VI that accepts a "path" string and something (a variant? Another JSON string?) describing the "complex object" and generates the JSON patch operation. This would eliminate the need to duplicate this logic across several VIs.
As in many cases, this is something that's straightforward to do in most languages but, at least for me, not in LabVIEW. I've posted a couple of snippets below to illustrate what I mean better.
In the example below, I want to create a JSON patch operation that says "add a new transmit message definition under "testcan.candb" with the name "example_tx_msg"":
In this example, I want to create another JSON patch operation that says "add a field to that message definition called "example_tx_field"":
These VIs do not work as intended, however, which makes sense because I'm passing the JSON representation of a complex object to "value" rather than the object itself. I expect to get JSON like this:
{
"op": "add",
"path": "testcan.candb.example_tx_msg",
"value": {
"type": "tx",
"messageId": 540,
"repeatRate": 100,
"canPort": 0,
"template": [255, 1]
}
}
I actually get JSON like this:
{
"op": "add",
"path": "testcan.candb.example_tx_msg",
"value": "{\"type\":\"tx\",\"messageId\":540,\"repeatRate\":100,\"canPort\":0,\"template\":[255,1]}"
}
I'd very strongly prefer a solution to this to use only native LabVIEW VI's. That's because:
- Unlike pretty much every other package management system, the de facto one for LabVIEW (VIPM) charges you to use half its features.
- In any case, VIPM fails to install any OpenG package for me (HTTP 400: bad request), and by extension any packages depending on these (such as JKI's JSON API package).
hi david!
the thing is, that when you try to create json from your cluster,
you already have the first as string, this is exactly what you get in your resulting JSON.
replace the "value" in second cluster with a variant,
then use JKI JSON library (from vipm), which lets you convert anything to json (the labview json converter, won't convert enums for example)
Hi jwscs,
the thing is, that when you try to create json from your cluster,
you already have the first as string, this is exactly what you get in your resulting JSON.
I'm aware, that's why I wrote "These VIs do not work as intended, however, which makes sense because I'm passing the JSON representation of a complex object to "value" rather than the object itself."
I was trying to convey the gist of what I wanted with a setup that I knew didn't work but was close enough.
replace the "value" in second cluster with a variant
This was the first thing I tried, but it doesn't seem to work with "Flatten To JSON"
then use JKI JSON library (from vipm), which lets you convert anything to json (the labview json converter, won't convert enums for example)
As I mentioned in my post, I was trying to avoid using third-party libraries for this. In particular, this is what happens when I try to install JKI's JSON library using VIPM 2017:
Still, I'm grateful you provided the solution you did. It's nice to know for sure that I can use JKI's JSON library if I ever get it installed.
