Updating an attribute: error I don't know how to fix

[ubacher]

I read an attribute and Im returns:

Code Select Expand

get attrib returns:{
  "result": [
    {
      "id": 750963,
      "data": [
        {
          "instanceId": 1,
          "PageSequenceNr": "HellO"
        }
      ]
    }
  ],
  "debug": {
    "runtime": "0ms"
  }
}

I then try to update the attribute and I pass the following to the server:

Code Select Expand

{"setid":1,"id":750963,"actions":"[{\"op\":\"update\",\"data\":{\"PageSequenceNr\":\"HellO\",\"instanceid\":1}}]"}

(If I specify add instead of update it works OK!

However I get the error message:

Code Select Expand

Set attrib returns:{
  "result": "failed",
  "resultMessage": "instanceid is missing or not an array. To delete all instances, specify an empty array.",
  "debug": {
    "runtime": "0ms"
  }
}

How can I fix this. The full script is attached.
( PS: I did start out with the attributes sample script - but it did not help me with this)

[Mario]

From the top of my mind: instanceid must be an array.

Wrong:

instanceid: 1

Correct:

instanceid: [1]

[ubacher]

I tried all kinds of things in terms of arrays but no luck.

[Mario]

Show me the JSON you are sending.

[ubacher]

JSON I am sending was shown in my first post. Here again:

Code Select Expand

{"setid":1,"id":750963,"actions":"[{\"op\":\"update\",\"data\":{\"PageSequenceNr\":\"HellO\",\"instanceid\":1}}]"}

[thrinn]

In your script you attached to the original post you seem to call the SetAttribute function with 'replace' for the AddOrUpdate parameter. Should this not be 'update'?

[Mario]

Your code uses

action.data.instanceid = 1;

but instanceid must be an array. Hence the "not an array" error message.,
Change to

action.data.instanceid = [1];


to make this an array (with one element in this case).

Note: Take care when you work with instance ids. These record numbers are to be considered volatile. Don't hard-code them in your scripts - except for testing purposes.

[ubacher]

Does not work:

Code Select Expand

This is sent to the server: {"setid":1,"id":750963,"actions":"[{\"op\":\"update\",\"data\":{\"PageSequenceNr\":\"Hello\",\"instanceid\":[1]}}]"}
myscript attributes read write.js:120 Set attrib returns:{
  "result": "failed",
  "resultMessage": "instanceid is missing or not an array. To update all instances, specify an empty array.",
  "debug": {
    "runtime": "0ms"
  }
}

[Mario]

I've made a check with the Attributes Samples App.

Under 2. Updating Data I selected a set and then changed the text for the instance 1 of the currently selected file. Works without problems.
Did you try that? Does it work?
The JSON record sent to the server looks like this:

actions:[{"op":"update","instanceid":["1"],"data":{"Note":"A nice baby photo"}}]

   "actions": [{
      "op": "update",
      "instanceid": [1],
      "data": {
         "Note": "This is the new content of Note"
      }
   }]

[ubacher]

Thank you - this info allowed me to find my error.

I had

Code Select Expand

action.data.instanceid = [1];

instead of

Code Select Expand

action.instanceid = [1];

What caused the confusion was that when getting the attribute it returns:

Code Select Expand

get attrib returns:{
  "result": [
    {
      "id": 748569,
      "data": [
        {
          "instanceId": 1,
          "Description": "hello",
          "PageSequenceNr": "000"
        }
      ]
    }
  ],
  "debug": {
    "runtime": "0ms"
  }
Where instanceId is under data. (Although this instanceid is spelled with a big I in Id.)

[Mario]

1. The data returned can be for multiple files and contain multiple instance ids (if a file has multiple rows of data), hence the instance id is always part of data.

2. As explained in the documentation (Naming Conversions), parameter names sent to IMWS are always lower case. Data returned in JSON format uses the standard camelCase syntax.

[thrinn]

2. As explained in the documentation (Naming Conversions), parameter names sent to IMWS are always lower case.

If only I would remember all that is written... 
I did have quite a look at ubacher's example coding, but I did not see this.

Am I correct in assuming that the lower case rule is only valid for "native" IMatch parameters, while e.g. attribute names have to be spelled exactly the way they have been created?

[Mario]

In JavaScript, everything is case-sensitive.

Since the IMWS endpoints are language-agnostic (work with all languages) and should also work with command-line tools like cUrl or wget (which need to code everything into the URL) I decided t make the endpoint names and parameters all lower-case.

By common standards, JavaScript code and, by extension, JSON data storage, should use camelCase names, starting with a lower letter. So you get instanceId, folderName, etc.

IMatch is also case-sensitive so the Category "family" is different from "Family". Same for Attributes. If you access a category or Attribute or file by path/name, always write them as they show up in IMatch itself. In case you do it wrong, you just get an "not found" back so you'll know.