GraphQL error in job execution

[raiAmagi]

I am learning about Dagster's I/O manager.
We are using the old documentation for learning.

I built a learning code with both Asset and Op functions based on the following document and was verifying its operation.
When I was testing the code, I got an “Unexpected GraphQL error” on the “Open launchpad” screen of the job.
Unexpected GraphQL error” occurred on the ‘Open launchpad’ screen of the job.

The situation is as follows:
　1. reloading the code location completes without error.
　2. execution of the following two jobs from the “Materialize all” button on the Jobs screen completes without error.
　　　　iomanager_basic_asset_job
　　　　custom_iomanager_basic_asset_job
　3. the following error occurs when “Open launchpad” is selected on the Jobs screen for the two jobs created.
　　　Error: Unexpected GraphQL error
　　　　　　　　　The details are described below.
　　 However
　　　The two jobs can be materialized by “Open launchpad”.
　　　Both jobs do not output any errors in the execution log.

As mentioned above, both Materialize and Open launchpad do not generate errors in execution,
I would like to know the cause of the error that appears when the “Open launchpad” screen is displayed.

※ The GraphQL error describes an error related to partitions.
　 Since we had tried and failed in the “Handling partition mappings” document just before, we thought this might be the cause of the error.
　 Re-creating the database schema just in case this was the cause,
　 However, the GraphQL error persisted.

The environment used is as follows.
　Python：　3.12.8
　Dagster:　1.9.12
　DB: PostgreSQL
　OS:　Windows11

Execution command:　dagster dev -f iomanager_learning.py

Documents being studied:
　Writing a per-input I/O manager
　https://legacy-docs.dagster.io/concepts/io-management/io-managers#writing-a-per-input-io-manager

---

Code built for learning

import os
import numpy as np
import pandas as pd
from dagster import (
asset,
AssetExecutionContext,
Definitions,
ConfigurableIOManager,
InputContext,
OutputContext,
define_asset_job,
In,
AssetIn,
)

-=-=-=-=-=-=-=-=-=-=-=-=

code01

-=-=-=-=-=-=-=-=-=-=-=-=

@asset
def iomanager_basic_upstream_asset():
return [1, 2, 3]

@asset
def iomanager_basic_downstream_asset(iomanager_basic_upstream_asset):
return iomanager_basic_upstream_asset + [4]

iomanager_basic_asset_job = define_asset_job(
description='code01',
name = 'iomanager_basic_asset_job',
selection=[
iomanager_basic_upstream_asset,
iomanager_basic_downstream_asset
],
)

-=-=-=-=-=-=-=-=-=-=-=-=

code02

-=-=-=-=-=-=-=-=-=-=-=-=

class PandasIOManager(ConfigurableIOManager):
path_prefix: str

def _get_path(self, context: OutputContext):
    return os.path.join(self.path_prefix, 'io_manager', f'{context.asset_key.path[-1]}.csv')

def handle_output(self, context: "OutputContext", obj: pd.DataFrame):
    if obj is not None:
        file_path = self._get_path(context)
        return obj.to_csv(file_path)

def load_input(self, context: "InputContext") -> pd.DataFrame:
    file_path = self._get_path(context)
    return pd.read_csv(file_path)

@asset(io_manager_key='pandas_manager')
def custom_iomanager_basic_upstream_asset() -> pd.DataFrame:
return pd.DataFrame([1, 2, 3])

@asset(io_manager_key='pandas_manager')
def custom_iomanager_basic_downstream_asset(context: AssetExecutionContext,
custom_iomanager_basic_upstream_asset: pd.DataFrame
) -> pd.DataFrame:
custom_iomanager_basic_upstream_asset.loc['ONE'] = 0
context.log.info(custom_iomanager_basic_upstream_asset)
return pd.DataFrame(custom_iomanager_basic_upstream_asset)

custom_iomanager_basic_asset_job = define_asset_job(
description='code02',
name = 'custom_iomanager_basic_asset_job',
selection=[
custom_iomanager_basic_upstream_asset,
custom_iomanager_basic_downstream_asset,
],

)

-=-=-=-=-=-=-=-=-=-=-=-=

DagsterDefinitions

-=-=-=-=-=-=-=-=-=-=-=-=

defs = Definitions(
assets = [
iomanager_basic_upstream_asset,
iomanager_basic_downstream_asset,

    custom_iomanager_basic_upstream_asset,
    custom_iomanager_basic_downstream_asset,
],
resources = {
    'pandas_manager': PandasIOManager(path_prefix='..\dagster_home\data'),
},
jobs = [
    iomanager_basic_asset_job,
    custom_iomanager_basic_asset_job,
],

)

---

Contents of Unexpected GraphQL error

Operation name: {ConfigPartitionsAssetsQuery}

Message: Cannot return null for non-nullable field PartitionKeys.partitionKeys.

Path: ["pipelineOrError","partitionKeysOrError","partitionKeys"]

Locations: [{"line":16,"column":11}]

Stack Trace:
File "D:\pg\workspase.venv.venv_dagster_latest\Lib\site-packages\graphql\execution\execute.py", line 542, in execute_field
completed = self.complete_value(
^^^^^^^^^^^^^^^^^^^^
File "D:\pg\workspase.venv.venv_dagster_latest\Lib\site-packages\graphql\execution\execute.py", line 622, in complete_value
raise TypeError(

[raiAmagi]

The problem was solved by performing the “Database migration” in the “
Migrating to 1.9.0” section of the version migration.
I had overlooked the fact that database migration was required when migrating to 1.9.x.