Torch Data Bigquery
Torch Dataset
interface for Google BigQuery datasets.
Table of Contents
Installation
pip install torch-data-bigquery
Examples
BigQueryStorageDataset
import torch
from torch_data_bigquery import BigQueryStorageDataset
dataset = BigQueryStorageDataset(
billing_project=PROJECT,
selected_fields=[
"PassengerId",
"Survived",
"Pclass",
],
location=f"bq://your-gcp-project.kaggle_titanic.train",
)
dataloader = torch.utils.data.DataLoader(dataset)
BigQueryDataset
import torch
from torch_data_bigquery import BigQueryDataset
dataset = BigQueryDataset(
billing_project=PROJECT,
query=f"""
SELECT
Survived AS survived, # 1
Pclass AS pclass, # 2
DENSE_RANK() OVER(ORDER by Sex) AS sex, # 3
COALESCE(Age, AVG(Age) OVER()) AS age, # 4
SibSp AS siblings, # 5
Parch AS parents, # 6
Fare AS fare, # 7
DENSE_RANK() OVER(ORDER by Embarked) AS embarked, # 8
FROM `your-gcp-project.kaggle_titanic.train`
""",
)
dataloader = torch.utils.data.DataLoader(dataset)
License
torch-data-bigquery
is distributed under the terms of the MIT license.