0

I've simple (or not) question. How I can set num_rows for synthetic_data generated by HMASynthesizer? Tables:

region_id address
0 r_0 Cohenville
1 r_1 Lake Martha
2 r_2 West Josephfurt
3 r_3 East Valerieshire
4 r_4 Madisonport
user_id names region_id
0 0 Tammy r_2
1 1 Edward r_1
2 2 Veronica r_3
3 3 Kelly r_1
4 4 Jennifer r_0 
import pandas as pd
from faker import Faker
from sdv.metadata import MultiTableMetadata
from sdv.multi_table import HMASynthesizer

fake = Faker('en_US)

multi_table_data = {
    'users': df,
    'regions': regions
}

metadata = MultiTableMetadata()

metadata.detect_table_from_dataframe(
    table_name='users',
    data=df
)

metadata.update_column(
    table_name='users',
    column_name='user_id',
    sdtype='id',
    regex_format='[0-9]{1}')


metadata.update_column(
    table_name='users',
    column_name='region_id',
    sdtype='id',
    regex_format='[A-Za-z]{3}')

metadata.update_column(
    table_name='users',
    column_name='names',
    sdtype='first_name',
)

metadata.set_primary_key(
    table_name='users',
    column_name='user_id'
)

metadata.detect_table_from_dataframe(
    table_name='regions',
    data=regions
)

metadata.update_column(
    table_name='regions',
    column_name='region_id',
    sdtype='id',
    regex_format='r_[0-9]{1}',
)

metadata.update_column(
    table_name='regions',
    column_name='address',
    sdtype='address',
)

metadata.set_primary_key(
    table_name='regions',
    column_name='region_id'
)

metadata.add_relationship(
    parent_table_name='regions',
    child_table_name='users',
    parent_primary_key='region_id',
    child_foreign_key='region_id'
)

synthesizer = HMASynthesizer(metadata, locales=['en_US'])
synthesizer.fit(multi_table_data)

synthetic_data = synthesizer.sample()

Synthetic data (I can't control size of dataframes):

user_id names region_id
0 0 Jennifer r_0
1 1 David r_1
2 2 Thomas r_1
3 3 Claudia r_2
4 4 Bruce r_3
5 5 Lisa r_3
6 6 Andrew r_4
region_id address
0 r_0 77809 Rush Mountain Suite 952 Garciaton, DC 32497
1 r_1 USS Morrow FPO AP 28106
2 r_2 123 Dennis Points Humphreymouth, IN 32470
3 r_3 Unit 2560 Box 7577 DPO AA 88965
4 r_4 46038 Karen Via Apt. 979 Arnoldmouth, PA 22955
John Doe
  • 95
  • 6

1 Answers1

0

Size of synthetic data can be set by scale: synthetic_data = synthesizer.sample(scale=0.2)

John Doe
  • 95
  • 6