Snowflake ID execution in ekuiper will lose precision

[echils]

Snowflake ID execution in ekuiper will lose precision

Environment:

• eKuiper version (e.g. 1.3.0): 1.14.0
• Hardware configuration (e.g. lscpu): x86_64
• OS (e.g. cat /etc/os-release): Debian GNU/Linux
• Others:

What happened and what you expected to happen:

fix it please!

How to reproduce it (as minimally and precisely as possible):

Enter a snowflake ID in the data source

Anything else we need to know?:

[ngjaying]

@echils Thanks for rasing this up. We are willing to fix any problems, but before that, we need to know the problem. How to reproduce it. I am not sure this one line explanation can help or not.

[echils]

My use case @ngjaying ：

1. Create a kafka source named root ，the sample data structures are shown below

    {
      "createTime": "2024-12-12T14:34:37.516Z",
      "data": {
          "abnormal": false,
          "data": [
          {
              "field": "unit",
              "tag": false,
              "value": "m/s"
          }
          ],
          "definitionId": 1795292668348461056,
          "time": "2024-12-12 14:34:37",
          "value": 0
      },
      "definitionId": 1795292668348461056,
      "messageId": "schedule-1867095692340801536",
      "msg": null,
      "status": 200
    }

2. Create a simple rule

 select data.definitionId from root

3. Create a mqtt sink
4. The mqtt result has lost accuracy
   

[echils]

Hi, is there any progress on this issue?

[Yisaer]

It's fixed in the #3445

You can try the latest master version and add useInt64ForWholeNumber: true in the source conf used by the target stream

[younes199511]

Hello @ngjaying @Yisaer

Is it possible to fix this bug in 1.14 version ?

Thank you