Open Data Blend for Python
Open Data Blend for Python is the fastest way to get data from the Open Data Blend Dataset API. It is a lightweight, easy-to-use extract and load (EL) tool.
You can use the get_data function to download any data file belonging to an Open Data Blend dataset. Alternatively, can use the get_data_files function to download a collection of data files from an Open Data Blend dataset. The functions transparently download and cache the data locally or in cloud storage, mirroring the same folder hierarchy as on the remote server. They also cache a copy of the dataset metadata file (datapackage.json) at the point that they are called. The cache is persistent which means the files will be kept until they are deleted.
The versioned dataset metadata can be used to re-download a specific version of a data file (sometimes referred to as 'time travel'). You can learn more about how we version our datasets in the Open Data Blend Docs.
In addition to downloading the data and metadata files, get_data returns an object called Output which includes the locations of the downloaded files. Similarly, get_data_files returns an object called OutputSet which includes the locations of the files that are downloaded and the associated metadata. From there, you can query and analyse the data directly using something light like Pandas or, for more resource intensive processing, a data lakehouse platform like Databricks, or a scalable in-memory OLAP library like Polars.
Installation
Install the latest version of opendatablend from PyPI:
pip install opendatablendUsage Examples
NOTE
If you want to run the examples, be sure to replace placeholder values such as <ACCESS_KEY> with appropriate string literals or variables.
Some of the following examples require the pandas and pyarrow packages to be installed:
pip install pandas
pip install pyarrowMaking Public API Requests
NOTE
Public API requests have a monthly limit.
Get the Data
import opendatablend as odb
import pandas as pd
dataset_path = 'https://packages.opendatablend.io/v1/open-data-blend-road-safety/datapackage.json'
# Specify the resource name of the data file. In this example, the 'date' data file will be requested in Parquet format.
resource_name = 'date-parquet'
# Get the data and store the output object
output = odb.get_data(dataset_path, resource_name)
# Print the file locations
print(output.data_file_name)
print(output.metadata_file_name)Use The Data
# Read a subset of the columns into a dataframe
df_date = pd.read_parquet(output.data_file_name, columns=['drv_date_key', 'drv_date', 'drv_month_name', 'drv_month_number', 'drv_quarter_name', 'drv_quarter_number', 'drv_year'])
# Check the contents of the dataframe
df_dateMaking Authenticated API Requests
Get the Data
import opendatablend as odb
import pandas as pd
dataset_path = 'https://packages.opendatablend.io/v1/open-data-blend-road-safety/datapackage.json'
access_key = '<ACCESS_KEY>'
# Specify the resource name of the data file. In this example, the 'date' data file will be requested in Parquet format.
resource_name = 'date-parquet'
# Get the data and store the output object
output = odb.get_data(dataset_path, resource_name, access_key=access_key)
# Print the file locations
print(output.data_file_name)
print(output.metadata_file_name)Use the Data
# Read a subset of the columns into a dataframe
df_date = pd.read_parquet(output.data_file_name, columns=['drv_date_key', 'drv_date', 'drv_month_name', 'drv_month_number', 'drv_quarter_name', 'drv_quarter_number', 'drv_year'])
# Check the contents of the dataframe
df_dateDownloading Multiple Data Files
The get_data_files function can be used to download a set of data files by providing their resource names as a list.
Get the Data
import opendatablend as odb
import pandas as pd
dataset_path = 'https://packages.opendatablend.io/v1/open-data-blend-road-safety/datapackage.json'
access_key = '<ACCESS_KEY>'
# Specify the resource names of the data files. In this example, a subset of the available data files will be requested in Parquet format.
resource_names = [
'date-parquet',
'time-of-day-parquet',
'geolocation-parquet',
'road-safety-accident-info-parquet',
'road-safety-accident-location-parquet',
'road-safety-accident-2021-parquet'
]
# Get the data files and store the output object
output = odb.get_data_files(dataset_path, resource_names, access_key=access_key)
# Print the file locations
print(output.data_file_names)
print(output.metadata_file_name)Ingesting Data Directly into Cloud Storage Services
Azure Blob Storage
Using get_data
import opendatablend as odb
dataset_path = 'https://packages.opendatablend.io/v1/open-data-blend-road-safety/datapackage.json'
access_key = '<ACCESS_KEY>' # The access key can be set to an empty string if you are making a public API request
# Specify the resource name of the data file. In this example, the 'date' data file will be requested in Parquet format.
resource_name = 'date-parquet'
# Get the data and store the output object using the Azure Blob Storage file system
configuration = {
"connection_string" : "DefaultEndpointsProtocol=https;AccountName=<AZURE_BLOB_STORAGE_ACCOUNT_NAME>;AccountKey=<AZURE_BLOB_STORAGE_ACCOUNT_KEY>;EndpointSuffix=core.windows.net",
"container_name" : "<AZURE_BLOB_STORAGE_CONTAINER_NAME>" # e.g. odbp-integration
}
output = odb.get_data(dataset_path, resource_name, access_key=access_key, file_system="azure_blob_storage", configuration=configuration)
# Print the file locations
print(output.data_file_name)
print(output.metadata_file_name)
Using get_data_files
import opendatablend as odb
dataset_path = 'https://packages.opendatablend.io/v1/open-data-blend-road-safety/datapackage.json'
access_key = '<ACCESS_KEY>' # The access key can be set to an empty string if you are making a public API request
# Specify the resource names of the data files. In this example, a subset of the available data files will be requested in Parquet format.
resource_names = [
'date-parquet',
'time-of-day-parquet',
'geolocation-parquet',
'road-safety-accident-info-parquet',
'road-safety-accident-location-parquet',
'road-safety-accident-2021-parquet'
]
# Get the data and store the output object using the Azure Blob Storage file system
configuration = {
"connection_string" : "DefaultEndpointsProtocol=https;AccountName=<AZURE_BLOB_STORAGE_ACCOUNT_NAME>;AccountKey=<AZURE_BLOB_STORAGE_ACCOUNT_KEY>;EndpointSuffix=core.windows.net",
"container_name" : "<AZURE_BLOB_STORAGE_CONTAINER_NAME>" # e.g. odbp-integration
}
output = odb.get_data_files(dataset_path, resource_names, access_key=access_key, file_system="azure_blob_storage", configuration=configuration)
# Print the file locations
print(output.data_file_names)
print(output.metadata_file_name)Azure Data Lake Storage (ADLS) Gen2
Using get_data
import opendatablend as odb
dataset_path = 'https://packages.opendatablend.io/v1/open-data-blend-road-safety/datapackage.json'
access_key = '<ACCESS_KEY>' # The access key can be set to an empty string if you are making a public API request
# Specify the resource name of the data file. In this example, the 'date' data file will be requested in Parquet format.
resource_name = 'date-parquet'
# Get the data and store the output object using the Azure Data Lake Storage Gen2 file system
configuration = {
"connection_string" : "DefaultEndpointsProtocol=https;AccountName=<ADLS_GEN2_ACCOUNT_NAME>;AccountKey=<ADLS_GEN2_ACCOUNT_KEY>;EndpointSuffix=core.windows.net",
"container_name" : "<ADLS_GEN2_CONTAINER_NAME>" # e.g. odbp-integration
}
output = odb.get_data(dataset_path, resource_name, access_key=access_key, file_system="azure_blob_storage", configuration=configuration)
# Print the file locations
print(output.data_file_name)
print(output.metadata_file_name)
Using get_data_files
import opendatablend as odb
dataset_path = 'https://packages.opendatablend.io/v1/open-data-blend-road-safety/datapackage.json'
access_key = '<ACCESS_KEY>' # The access key can be set to an empty string if you are making a public API request
# Specify the resource names of the data files. In this example, a subset of the available data files will be requested in Parquet format.
resource_names = [
'date-parquet',
'time-of-day-parquet',
'geolocation-parquet',
'road-safety-accident-info-parquet',
'road-safety-accident-location-parquet',
'road-safety-accident-2021-parquet'
]
# Get the data and store the output object using the Azure Data Lake Storage Gen2 file system
configuration = {
"connection_string" : "DefaultEndpointsProtocol=https;AccountName=<ADLS_GEN2_ACCOUNT_NAME>;AccountKey=<ADLS_GEN2_ACCOUNT_KEY>;EndpointSuffix=core.windows.net",
"container_name" : "<ADLS_GEN2_CONTAINER_NAME>" # e.g. odbp-integration
}
output = odb.get_data_files(dataset_path, resource_names, access_key=access_key, file_system="azure_blob_storage", configuration=configuration)
# Print the file locations
print(output.data_file_names)
print(output.metadata_file_name)Amazon S3
Using get_data
import opendatablend as odb
dataset_path = 'https://packages.opendatablend.io/v1/open-data-blend-road-safety/datapackage.json'
access_key = '<ACCESS_KEY>' # The access key can be set to an empty string if you are making a public API request
# Specify the resource name of the data file. In this example, the 'date' data file will be requested in Parquet format.
resource_name = 'date-parquet'
# Get the data and store the output object using the Amazon S3 file system
configuration = {
"aws_access_key_id" : "<AWS_ACCESS_KEY_ID>",
"aws_secret_access_key" : "AWS_SECRET_ACCESS_KEY",
"bucket_name" : "<BUCKET_NAME>", # e.g. odbp-integration
"bucket_region" : "<BUCKET_REGION>" # e.g. eu-west-2
}
output = odb.get_data(dataset_path, resource_name, access_key=access_key, file_system="amazon_s3", configuration=configuration)
# Print the file locations
print(output.data_file_name)
print(output.metadata_file_name)
Using get_data_files
import opendatablend as odb
dataset_path = 'https://packages.opendatablend.io/v1/open-data-blend-road-safety/datapackage.json'
access_key = '<ACCESS_KEY>' # The access key can be set to an empty string if you are making a public API request
# Specify the resource names of the data files. In this example, a subset of the available data files will be requested in Parquet format.
resource_names = [
'date-parquet',
'time-of-day-parquet',
'geolocation-parquet',
'road-safety-accident-info-parquet',
'road-safety-accident-location-parquet',
'road-safety-accident-2021-parquet'
]
# Get the data and store the output object using the Amazon S3 file system
configuration = {
"aws_access_key_id" : "<AWS_ACCESS_KEY_ID>",
"aws_secret_access_key" : "AWS_SECRET_ACCESS_KEY",
"bucket_name" : "<BUCKET_NAME>", # e.g. odbp-integration
"bucket_region" : "<BUCKET_REGION>" # e.g. eu-west-2
}
output = odb.get_data(dataset_path, resource_names, access_key=access_key, file_system="amazon_s3", configuration=configuration)
# Print the file locations
print(output.data_file_names)
print(output.metadata_file_name)Google Cloud Storage
Using get_data
import opendatablend as odb
dataset_path = 'https://packages.opendatablend.io/v1/open-data-blend-road-safety/datapackage.json'
access_key = '<ACCESS_KEY>' # The access key can be set to an empty string if you are making a public API request
# Specify the resource name of the data file. In this example, the 'date' data file will be requested in Parquet format.
resource_name = 'date-parquet'
# Get the data and store the output object using the Google Cloud Storage file system
configuration = {
"service_account_private_key_file" : "<PATH_TO_SERVICE_ACCOUNT_PRIVATE_KEY_FILE>",
"bucket_name" : "<BUCKET_NAME>", # e.g. odbp-integration
"bucket_location" : "<BUCKET_LOCATION>" # e.g. europe-west2
}
output = odb.get_data(dataset_path, resource_name, access_key=access_key, file_system="google_cloud_storage", configuration=configuration)
# Print the file locations
print(output.data_file_name)
print(output.metadata_file_name)
Using get_data_files
import opendatablend as odb
dataset_path = 'https://packages.opendatablend.io/v1/open-data-blend-road-safety/datapackage.json'
access_key = '<ACCESS_KEY>' # The access key can be set to an empty string if you are making a public API request
# Specify the resource names of the data files. In this example, a subset of the available data files will be requested in Parquet format.
resource_names = [
'date-parquet',
'time-of-day-parquet',
'geolocation-parquet',
'road-safety-accident-info-parquet',
'road-safety-accident-location-parquet',
'road-safety-accident-2021-parquet'
]
# Get the data and store the output object using the Google Cloud Storage file system
configuration = {
"service_account_private_key_file" : "<PATH_TO_SERVICE_ACCOUNT_PRIVATE_KEY_FILE>",
"bucket_name" : "<BUCKET_NAME>", # e.g. odbp-integration
"bucket_location" : "<BUCKET_LOCATION>" # e.g. europe-west2
}
output = odb.get_data(dataset_path, resource_names, access_key=access_key, file_system="google_cloud_storage", configuration=configuration)
# Print the file locations
print(output.data_file_names)
print(output.metadata_file_name)OneLake in Microsoft Fabric
You can use Open Data Blend for Python to ingest data directly into OneLake in Microsoft Fabric using a Fabric Notebook.
Prerequisites
Before attempting to ingest the data into OneLake using this method, you need to:
- Create a Microsoft Fabric Lakehouse
- Create a Microsoft Fabric Notebook
- Set a default lakehouse for your Microsoft Fabric Notebook
- Install the 'opendatablend' library through the Library Management section of the workspace settings
You can then use the following methods to ingest the data. Pay special attention to base_path value because this is what controls where the data will be stored within OneLake. The base_path must point to the 'Files' location or a subfolder within it.
Using get_data
import opendatablend as odb
dataset_path = 'https://packages.opendatablend.io/v1/open-data-blend-road-safety/datapackage.json'
access_key = '<ACCESS_KEY>' # The access key can be set to an empty string if you are making a public API request
# Specify the resource name of the data file. In this example, the 'date' data file will be requested in Parquet format.
resource_name = 'date-parquet'
# Specify the base path for where the data files should be landed. In this example, we want them to be stored in the root of the 'Files' folder in the Fabric lakehouse that has been attached to the Fabric notebook.
base_path = '/lakehouse/default/Files/'
output = odb.get_data(dataset_path, resource_name, base_path=base_path, access_key=access_key)
# Print the file locations
print(output.data_file_name)
print(output.metadata_file_name)
Using get_data_files
import opendatablend as odb
dataset_path = 'https://packages.opendatablend.io/v1/open-data-blend-road-safety/datapackage.json'
access_key = '<ACCESS_KEY>' # The access key can be set to an empty string if you are making a public API request
# Specify the resource names of the data files. In this example, a subset of the available data files will be requested in Parquet format.
resource_names = [
'date-parquet',
'time-of-day-parquet',
'geolocation-parquet',
'road-safety-accident-info-parquet',
'road-safety-accident-location-parquet',
'road-safety-accident-2021-parquet'
]
# Specify the base path for where the data files should be landed. In this example, we want them to be stored in the root of the 'Files' folder in the Fabric lakehouse that has been attached to the Fabric notebook.
base_path = '/lakehouse/default/Files/'
output = odb.get_data_files(dataset_path, resource_names, base_path=base_path, access_key=access_key)
# Print the file locations
print(output.data_file_names)
print(output.metadata_file_name)Additional Examples
For more in-depth examples, see the examples folder.
