Bring your towel and join one of the fastest growing data communities. Welcome to our second-generation open source orchestration platform, a completely rethought approach to dataflow automation.

Prefect Community

Hi, I'm trying to build out a registration process that uses `flow.serialized_hash()` to only register the modified flows. However, `flow.serialized_hash()` is not returning a consistent hash value for flows that have not been modified. Why would this be the case?

For context, I'm testing by registering the flows locally with `flow.register()`

Hey <@U025TTA33FT>, just to make sure, you’re doing something like `flow.register(…, idempotency_key=flow.serialized_hash())` , or are you just running `flow.serialized_hash()` multiple times and comparing the values?

<@U01QEJ9PP53> I initially was trying `flow.register(…, idempotency_key=flow.serialized_hash())` but that wasn't working for me, so then I tried to just test `flow.serialized_hash()` by running it and comparing the values on independent registration runs.

I also compared the `flow.serialize()` values on independent runs and those are different values as well.

Will ask the team more about this and get back to you

After doing a little more digging, it looks like the difference in `flow.serialize()` is the following:

Test 1: `"start_date": {"dt": "2021-08-03T10:47:33.780910"`

Test2: `"start_date": {"dt": "2021-08-03T10:49:13.281844"`

Does the time of registration impact the `flow.serialize()` return value?

<@U01QEJ9PP53> Here's the entire json key/value for reference: `"clocks": [{"__version__": "0.15.1", "cron": "20 6 * * *", "day_or": true, "end_date": null, "labels": null, "parameter_defaults": {}, "start_date": {"dt": "2021-08-03T10:49:13.281844", "tz": "America/Chicago"}, "type": "CronClock"}],`

Can you give me the code for the Flow block? I guess it good if the clock start time if being moved. I’ll try this myself.

`start_date = pendulum.now(tz="America/Chicago")`
`clocks = [CronClock(s, start_date=start_date) for s in cron_strings]`

`run_schedule = create_cron_schedule("20 6 * * *")`

`with Flow(`
    `name="netchb-to-snowflake",`
    `schedule=netchb_run_schedule,`
`) as flow:`
   `flow.chain(task chain...)`

I think I understand the problem now. We are using the optional `start_date` parameter when building the Cron schedule. The `start_date` value is then used in the `flow.serialize()` json object which in turn causes inconsistent hash values because we are using `pendulum.now(tz="America/Chicago")` as the `start_date`.

I see. I’ll ask the team about how to approach this.

Hey <@U025TTA33FT>, so this is expected behavior, otherwise we would not be able to re-register flows that have start dates changes. Something you can to in this the `start_date` to the first of the year for something in the past so that the `serialized_hash` will stay the same.