Implementation:Sdv dev SDV BaseMultiTableSynthesizer Sample

Knowledge Sources	SDV SDV Documentation
Domains	Synthetic_Data, Relational_Data
Last Updated	2026-02-14 00:00 GMT

Overview

Concrete tool for generating synthetic multi-table relational data from a fitted synthesizer, provided by the SDV library.

Description

The BaseMultiTableSynthesizer.sample method generates synthetic data for all tables using hierarchical sampling. It validates that the synthesizer is fitted, applies the scale factor, invokes the hierarchical sampler, and returns a dictionary of DataFrames.

Usage

Call this method after fitting an HMASynthesizer. It returns a dictionary mapping table names to synthetic DataFrames.

Code Reference

Source Location

Repository: SDV
File: sdv/multi_table/base.py
Lines: L688-741

Signature

def sample(self, scale=1.0):
    """Generate synthetic data for the entire dataset.

    Args:
        scale (float): Scale factor for output row counts.
            1.0 = same as original, >1.0 = more rows, <1.0 = fewer rows.
            Defaults to 1.0.

    Returns:
        dict[str, pd.DataFrame]: Synthetic data for all tables.
    """

Import

from sdv.multi_table import HMASynthesizer
# sample is called as: synthesizer.sample(scale=1.0)

I/O Contract

Inputs

Name	Type	Required	Description
scale	float	No	Scale factor for row counts (default: 1.0, must be > 0)

Outputs

Name	Type	Description
return value	dict[str, pd.DataFrame]	Synthetic data for all tables with referential integrity

Usage Examples

# After fitting
synthetic_data = synthesizer.sample(scale=1.0)

for table_name, df in synthetic_data.items():
    print(f"{table_name}: {len(df)} rows")

# Generate 2x the original data
larger_data = synthesizer.sample(scale=2.0)

Related Pages

Page Connections

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

Principle

Implementation

Heuristic

Environment