Migrate Datasets Client

Migrate dataset management from ArizeDatasetsClient to ArizeClient.datasets in the v8 SDK.

This section covers migrating dataset management methods from v7’s ArizeDatasetsClient to v8’s ArizeClient.datasets.

Version 7

from arize.experimental.datasets import ArizeDatasetsClient

# v7 api_key parameter took developer key values
client = ArizeDatasetsClient(
    api_key="your-developer-key"  # Developer key (deprecated)
)

list_datasets()

The list_datasets() method migrates from client.list_datasets() to client.datasets.list().

Parameter Reference

Parameter	v7	v8	Changes
space_id	Required	Optional	Now optional; if not provided, lists datasets across all spaces
limit	N/A	✅ Optional	Maximum number of datasets to return (default 100)
cursor	N/A	✅ Optional	Pagination cursor for retrieving next page

Side-by-Side Comparison

Version 7

from arize.experimental.datasets import ArizeDatasetsClient

# Client initialization
client = ArizeDatasetsClient(api_key="your-developer-key")

# List datasets
datasets_df = client.list_datasets(space_id="your-space-id")

create_dataset()

The create_dataset() method migrates from client.create_dataset() to client.datasets.create().

Parameter Reference

Parameter	v7	v8	Changes
space_id	Required	Required	—
dataset_name	Required	Required	Renamed to name
name	N/A	✅ Required	Renamed from dataset_name
dataset_type	Required	❌ Removed	No longer required
data	Required	Required	Renamed to examples
examples	N/A	✅ Required	Renamed from data; accepts DataFrame or list of dicts
convert_dict_to_json	Optional	❌ Removed	Automatic conversion in v8
max_chunk_size	Optional	❌ Removed	Now configured at client level
force_http	N/A	✅ Optional	Force HTTP upload instead of gRPC (default False)

Side-by-Side Comparison

Version 7

from arize.experimental.datasets import ArizeDatasetsClient
from arize.pandas.proto import flight_pb2
import pandas as pd

# Client initialization
client = ArizeDatasetsClient(api_key="your-developer-key")

# Create dataset
dataset_id = client.create_dataset(
    space_id="your-space-id",
    dataset_name="my-dataset",
    dataset_type=flight_pb2.DatasetType.GENERATIVE,
    data=dataset_df,
    convert_dict_to_json=True,
    max_chunk_size=1000
)

get_dataset()

The get_dataset() method has a different behavior in v8. In v7, client.get_dataset() returned the dataset examples (underlying data). In v8, client.datasets.get() returns only the dataset metadata and versions, while client.datasets.list_examples() retrieves the actual examples.

Parameter Reference

For dataset metadata (v8’s datasets.get()):

Parameter	v7	v8	Changes
space_id	Required	❌ Removed	Not needed in v8
dataset_id	Optional	Required	Now required; no longer accepts dataset_name
dataset_name	Optional	❌ Removed	Use dataset_id instead
dataset_version	Optional	❌ Removed	All versions are returned in metadata
convert_json_str_to_dict	Optional	N/A	Only applies to examples, not metadata

For dataset examples (v8’s datasets.list_examples()):

Parameter	v7	v8	Changes
dataset_id	Optional	Required	—
dataset_version	Optional	Optional	Renamed to dataset_version_id
dataset_version_id	N/A	✅ Optional	If empty, returns latest version
limit	N/A	✅ Optional	Maximum number of examples per page (default 100); ignored if all=True
all	N/A	✅ Optional	When True, retrieves all examples via Flight (bypasses pagination). When False (default), uses REST with pagination
convert_json_str_to_dict	Optional	❌ Removed	Automatic conversion in v8

Side-by-Side Comparison

Version 7

from arize.experimental.datasets import ArizeDatasetsClient

# Client initialization
client = ArizeDatasetsClient(api_key="your-developer-key")

# Get dataset examples (underlying data) by ID
dataset_df = client.get_dataset(
    space_id="your-space-id",
    dataset_id="dataset-123",
    dataset_version="v1",
    convert_json_str_to_dict=True
)
# Returns: pandas DataFrame with the dataset examples

# Or get by name
dataset_df = client.get_dataset(
    space_id="your-space-id",
    dataset_name="my-dataset"
)

delete_dataset()

The delete_dataset() method migrates from client.delete_dataset() to client.datasets.delete().

Parameter Reference

Parameter	v7	v8	Changes
space_id	Required	❌ Removed	Not needed in v8
dataset_id	Optional	Required	Now required; no longer accepts dataset_name
dataset_name	Optional	❌ Removed	Use dataset_id instead

Side-by-Side Comparison

Version 7

from arize.experimental.datasets import ArizeDatasetsClient

# Client initialization
client = ArizeDatasetsClient(api_key="your-developer-key")

# Delete dataset by ID
success = client.delete_dataset(
    space_id="your-space-id",
    dataset_id="dataset-123"
)

# Or delete by name
success = client.delete_dataset(
    space_id="your-space-id",
    dataset_name="my-dataset"
)