Implementation:Online ml River Stream Iter Array

Knowledge Sources	Online_ml_River
Domains	Online_Learning, Data_Streaming, Array_Processing
Last Updated	2026-02-08 16:00 GMT

Overview

Converts NumPy arrays or Python lists into a River-compatible data stream for online learning.

Description

The iter_array function transforms 2D arrays of features and 1D arrays of targets into an iterable stream of (features_dict, target) tuples. It supports feature naming, shuffling, multi-output targets, and can handle both NumPy arrays and Python lists. Features can be automatically labeled with integers or custom names can be provided.

Usage

Use this when you have data in array format (NumPy or lists) and want to convert it to River's dict-based format for online learning. It's essential for bridging batch learning datasets with River's streaming API.

Code Reference

Source Location

Repository: Online_ml_River
File: river/stream/iter_array.py

Signature

def iter_array(
    X: np.ndarray,
    y: np.ndarray | None = None,
    feature_names: list[base.typing.FeatureName] | None = None,
    target_names: list[base.typing.FeatureName] | None = None,
    shuffle: bool = False,
    seed: int | None = None,
) -> base.typing.Stream:
    ...

Import

from river import stream

I/O Contract

Parameter	Type	Description
X	np.ndarray	2D array of features or 1D array of strings
y	np.ndarray or None	Array of targets (1D or 2D for multi-output)
feature_names	list or None	Names for features (uses integers if None)
target_names	list or None	Names for multi-output targets
shuffle	bool	Whether to shuffle data before streaming
seed	int or None	Random seed for shuffling

Returns:

Type	Description
Iterator[(dict, Any)]	Stream of (features dict, target) tuples

Usage Examples

import numpy as np
from river import stream

# Example 1: Basic usage with feature names
X = np.array([[1, 2, 3], [11, 12, 13]])
y = np.array([True, False])

dataset = stream.iter_array(
    X, y,
    feature_names=['x1', 'x2', 'x3']
)

for x, target in dataset:
    print(f"Features: {x}, Target: {target}")
# Output:
# Features: {'x1': 1, 'x2': 2, 'x3': 3}, Target: True
# Features: {'x1': 11, 'x2': 12, 'x3': 13}, Target: False

# Example 2: With shuffling
X = np.array([[1, 2], [3, 4], [5, 6], [7, 8]])
y = np.array([10, 20, 30, 40])

shuffled_stream = stream.iter_array(
    X, y,
    feature_names=['a', 'b'],
    shuffle=True,
    seed=42
)

print("Shuffled order:")
for x, target in shuffled_stream:
    print(f"{x} -> {target}")

# Example 3: Text data
X_text = np.array(["hello world", "machine learning"])
y_text = np.array([1, 0])

for text, label in stream.iter_array(X_text, y_text):
    print(f"Text: {text}, Label: {label}")

# Example 4: Multi-output targets
X_multi = np.array([[1, 2], [3, 4]])
y_multi = np.array([[0.5, 0.3], [0.7, 0.2]])

multi_stream = stream.iter_array(
    X_multi, y_multi,
    feature_names=['f1', 'f2'],
    target_names=['out1', 'out2']
)

for x, y_dict in multi_stream:
    print(f"Features: {x}")
    print(f"Targets: {y_dict}")

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment