AttributeError when detecting multi table metadata from a CSV

[Raghuramrwttw]

im getting error like
AttributeError: 'SingleTableMetadata' object has no attribute '_load_data_from_csv'

this is my code:
metadata.detect_table_from_csv(
table_name='hotels',
filepath='C:/Users/Admin/Desktop/pythonsdv/hotels.csv'
)
and this is the error im getting
AttributeError Traceback (most recent call last)
Cell In[4], line 1
----> 1 metadata.detect_table_from_csv(
2 table_name='hotels',
3 filepath='C:/Users/Admin/Desktop/pythonsdv/hotels.csv'
4 )

File c:\Users\Admin\Desktop\pythonsdv\env\lib\site-packages\sdv\metadata\multi_table.py:357, in MultiTableMetadata.detect_table_from_csv(self, table_name, filepath)
355 self._validate_table_not_detected(table_name)
356 table = SingleTableMetadata()
--> 357 data = table._load_data_from_csv(filepath)
358 print(data)
359 table._detect_columns(data)

AttributeError: 'SingleTableMetadata' object has no attribute '_load_data_from_csv'

please help to get out from this error

[npatki]

Hi @Raghuramrwttw thanks for filing this issue. I'm going to classify this as a bug since I'm able to replicate it with the demo data.

Workaround

A simple workaround is to load the data yourself as a pandas dataframe. Then you can use the other function detect_table_from_dataframe.

Note that pandas is already pre-installed when you install SDV.

import pandas as pd

hotels_dataframe = pd.read_csv('C:/Users/Admin/Desktop/pythonsdv/hotels.csv')
metadata.detect_table_from_dataframe(
    table_name='hotels',
    data=hotels_dataframe
)

You can do this independently for each table in your database.

For SDV Developers

Seems like table._load_data_from_csv was left there accidentally. Shouldn't we be using the function from utils instead?

from sdv.utils import load_data_from_csv

It would also be good to add an integration test for this.

[amontanez24]

closed by #1533