A library to standardize your data operations

Databand’s open source library helps teams establish their DataOps lifecycle. Track pipeline metadata, debug workflows, and add dynamic automation to data processes.

Github Documentation

New features

We built our custom operators to provide out-of-the-box functionality for pipeline tracking and optimization. Click the Github link to learn more about the operator and how to integrate in your pipelines.

Apache Spark Run Operator

Track Redshift query metadata, performance metrics, and data profiles

github

Snowflake Metadata Operator

Track Snowflake query metadata and performance metrics.

github

DataFrame Profiling

Report data statistics and distributions on Pandas and Spark dataframes.

github

Pipeline logging and tracking

Store metadata to always be on top of your pipeline and data health.

Custom Metrics

Define and report any custom metric about your data every time your pipeline runs.

Data Profiles

Automatically generate data profiling and statistics on data files and tables.

Input / Output

Track workflow inputs and outputs and lineage of data across tasks and broader pipelines.

Documentation

from dbnd_snowflake.airflow_operators import LogSnowflakeResourceOperator, 
log_snowflake_resource_operator

select_query = 'select * from "SNOWFLAKE_SAMPLE_DATA"."TPCDS_SF100TCL".
"CUSTOMER" limit 1000'

# Airflow snowflake_operator
get_customers_task = SnowflakeOperator(
    sql=select_query,
    snowflake_conn_id="test_snowflake_conn",
    task_id="get_customers",
)

# Databand operator for Snowflake resource tracking
log_snowflake_resources_task = LogSnowflakeResourceOperator(
    query_text=select_query,
    snowflake_conn_id="airflow_snowflake_conn",
    warehouse=None,
    database=database,
    schema=schema,
    account=account,
    task_id="log_snowflake_resources_task",
)

from dbnd_snowflake import log_snowflake_resource_usage

log_snowflake_resource_usage(
    query_text,
    database="DATABASE",
    user="user",
    connection_string="snowflake://<user>:<password>@<account>/",
    session_id=123456,
)

Integration points

Databand Open Source contains three methods of integration with your data workflows.

Connectors

Plugins to connect with services like Apache Airflow, Azkaban, Deequ, MLFlow, and more.

Operators

Off-the-shelf tasks and templates for tracking metadata and enhanced code deployment.

Instrumentation

Logging methods and code annotations for reporting metrics and building dynamic workflows.

Optimized run orchestration

Leverage Databand operators and instrumentation to make building, running, and deploying pipelines easier and more dynamic. You can use the library directly in Python workflows from scratch, or as an extension of Apache Airflow.

Dynamic Runs

Define different versions of pipelines to run based on changing data or parameters

Central Configs

Abstract out configurations to compute environments and data locations

Decorator Definition

Easily define flows by annotating your code, without making big changes to your workflow

Documentation

@task
def buy_vegetables(veg_list)
	from store import veg_store
  	return veg_store.purchase(veg_list)

@task
def cut(vegetables):
    chopped = []
    for veg in vegetables:
        chopped.append(veg.dice())
    return [x + "\n" for x in chopped]

@task
def add_dressing(chopped_vegetables, dressing, salt_amount="low"):
    for veg in chopped_vegetables:
      	veg.season(salt_amount)
    return chopped_vegetables

@pipeline
def prepare_salad(vegetables_list=data_repo.vegetables, dressing="oil"):
	vegetables = buy_vegetables(vegetables_list)
    chopped = cut(vegetables)
	dressed = add_dressing(chopped, dressing)
	return dressed

with DAG(dag_id="prepare_salad") as dag:
	salad = prepare_salad()

""" CLI:
airflow backfill -s 2020-06-01 -e 2020-06-02 prepare_salad
"""

@task
def buy_vegetables(veg_list)
	from store import veg_store
  	return veg_store.purchase(veg_list)

@task
def cut(vegetables):
    chopped = []
    for veg in vegetables:
        chopped.append(veg.dice())
    return [x + "\n" for x in chopped]

@task
def add_dressing(chopped_vegetables, dressing, salt_amount="low"):
    for veg in chopped_vegetables:
      	veg.season(salt_amount)
    return chopped_vegetables

@pipeline
def prepare_salad(vegetables_list=data_repo.vegetables, dressing="oil"):
	vegetables = buy_vegetables(vegetables_list)
    chopped = cut(vegetables)
	dressed = add_dressing(chopped, dressing)
	return dressed

""" CLI:
dbnd run prepare_salad
"""

Start a free trial or demo

Free Trial REQUEST A DEMO