Hey, I'm looking for a way to change to description of a scheme with the python emitter. Is there a object that I can use for this like

DatasetPropertiesClass

Or can I use this object to also change the scheme of a dataset?

https://demo.datahubproject.io/dataset/urn:li:dataset:(urn:li:dataPlatform:datahub,Dataset,PROD)/Schema?is_lineage_mode=false are the components that makes up a dataset. to change the schema (for instance, add field/remove field), you should be ingesting a newer version of schemaMetaData personally, i don't like the way the objects are shown as part of a dataset... but schema_classes.py is no longer commited to the repo.

Hey @better-orange-49102 Thanks for the response.I'm not understanding how I should approach editing the meta data with python based on your comment, or maybe my question was wrong. https://datahubproject.io/docs/metadata-ingestion/as-a-library/ That link gives an example to change properties on a datasets. I want to programmatic edit the description of an field. For example in the picture I want to change the description of urn. How would I approach this problem?

something like

import datahub.emitter.mce_builder as builder
from datahub.emitter.mcp import MetadataChangeProposalWrapper
from datahub.metadata.schema_classes import ChangeTypeClass, EditableSchemaMetadataClass, EditableSchemaFieldInfoClass

from datahub.emitter.rest_emitter import DatahubRestEmitter

# Create an emitter to DataHub over REST
emitter = DatahubRestEmitter(gms_server="<http://localhost:8080>", extra_headers={})

# Test the connection
emitter.test_connection()

# Construct a dataset properties object
dataset_schema = EditableSchemaMetadataClass(editableSchemaFieldInfo=[
        EditableSchemaFieldInfoClass(fieldPath = 'colA', description='this is the desc for A'),
        EditableSchemaFieldInfoClass(fieldPath = 'colB', description='this is the desc for B'),
    ],
    )

# Construct a MetadataChangeProposalWrapper object.
metadata_event = MetadataChangeProposalWrapper(
    entityType="dataset",
    changeType=ChangeTypeClass.UPSERT,
    entityUrn=builder.make_dataset_urn("bigquery", "my-project.my-dataset.user-table"),
    aspectName="editableSchemaMetadata",
    aspect=dataset_schema,
)

# Emit metadata! This is a blocking call
emitter.emit(metadata_event)

Massive thanks! really appreciate the help!

however, this will overwrite any other ui edits to other fields that you didn't specify

for instance, if there are 3 fields in the dataset, and all 3 of them already have descriptions inputted via UI, now when you emit this, the 3th field's description will disappear.

Good to know! Thanks in our current setup we want to document everything in yaml files that correspond a dataset. So that would be perfect then we have one single source of truth which we can track in detail