Using Delta Lake with Daft
Daft is a framework for ETL, analytics, and ML/AI at scale with a familiar Python dataframe API, implemented in Rust.
For Delta Lake users, Daft is a great data processing tool because it gives you the following features:
- Skipping Filtered Data: Daft implements automatic partition pruning and stats-based file pruning for filter predicates, skipping data that doesn’t need to be read.
- Multimodal Dataframes: read, write and transform multimodal data incl. images, JSON, PDF, audio, etc.
- Parallel + Distributed Reads: Daft parallelizes Delta Lake table reads over all cores of your machine, if using the default multithreading runner, or all cores + machines of your Ray cluster, if using the distributed Ray runner.
- Multi-cloud Support: Daft supports reading Delta Lake tables from AWS S3, Azure Blob Store, and GCS, as well as local files.
Daft and Delta Lake work really well together. Daft provides unified compute for Delta Lake’s unified storage. Together, Delta Lake and Daft give you high-performance query optimization and distributed compute on massive datasets.
Installing Daft for Delta Lake
The easiest way to use Delta Lake format with Daft DataFrames is to install Daft with the [deltalake] extras using pip:
This adds the deltalake Python package to your install. This package is used to fetch metadata about the Delta Lake table, such as paths to the underlying Parquet files and table statistics. You can of course also install the deltalake manually.
Read Delta Lake into a Daft DataFrame
You can easily read Delta Lake tables into a Daft DataFrame using the read_delta_lake method. Let's use it to read some data stored in a Delta Lake on disk. You can access the data stored as a Delta Lake on Github
You can also read in Delta Lake data from remote sources like S3:
# table_uri = (
# "s3://daft-public-datasets/red-pajamas/"
# "stackexchange-sample-north-germanic-deltalake"
# )
# df = daft.read_delta_lake(table_uri)
| first_name Utf8 | last_name Utf8 | country Utf8 | continent Utf8 |
|---|
Daft DataFrames are lazy by default. This means that the contents will not be computed ("materialized") unless you explicitly tell Daft to do so. This is best practice for working with larger-than-memory datasets and parallel/distributed architectures.
The Delta table we have just loaded only has 5 rows. You can materialize it in memory using the .collect method:
> df.collect()
| | first_name | last_name | country | continent |
|---:|:-------------|:------------|:----------|:------------|
| 0 | Ernesto | Guevara | Argentina | NaN |
| 1 | Bruce | Lee | China | Asia |
| 2 | Jack | Ma | China | Asia |
| 3 | Wolfgang | Manche | Germany | NaN |
| 4 | Soraya | Jala | Germany | NaN |
What can I do with a Daft DataFrame?
Daft gives you full-featured DataFrame functionality, similar to what you might be used to from pandas, Dask or PySpark.
On top of this, Daft also gives you:
- Multimodal data type support to work with Images, URLs, Tensors and more
- Expressions API for easy column transformations
- UDFs for multi-column transformation, incl. ML applications
Let's take a quick look at some of Daft's basic DataFrame operations.
You can select columns from your DataFrame using the select method. We'll use the show method to show the first n rows (defaults to 10):
> df.select("first_name", "country").show()
| | first_name | country |
|---:|:-------------|:----------|
| 0 | Ernesto | Argentina |
| 1 | Bruce | China |
| 2 | Jack | China |
| 3 | Wolfgang | Germany |
| 4 | Soraya | Germany |
You can sort your Daft DataFrame using the sort method:
> df.sort(df["country"], desc=True).show()
| | first_name | last_name | country | continent |
|---:|:-------------|:------------|:----------|:------------|
| 0 | Wolfgang | Manche | Germany | NaN |
| 1 | Soraya | Jala | Germany | NaN |
| 2 | Bruce | Lee | China | Asia |
| 3 | Jack | Ma | China | Asia |
| 4 | Ernesto | Guevara | Argentina | NaN |
You can filter your DataFrame using the where method:
> df.where(df["continent"] == "Asia").show()
| | first_name | last_name | country | continent |
|---:|:-------------|:------------|:----------|:------------|
| 0 | Bruce | Lee | China | Asia |
| 1 | Jack | Ma | China | Asia |
You can group your DataFrame by a specific columns using the groupby method. You can then specify the aggregation method, in this case using the count aggregator:
> df.select("first_name", "country").groupby(df["country"]).count("first_name").show()
| | country | first_name |
|---:|:----------|-------------:|
| 0 | Germany | 2 |
| 1 | China | 2 |
| 2 | Argentina | 1 |
Check out the Daft User Guide for a complete list of DataFrame operations.
Data Skipping Optimizations
You may have noticed the Delta Lake warning at the top when we first called collect on our DataFrame:
Delta Lake is informing us that the data is partitioned on the country column.
Daft's native query optimizer has access to all of the Delta Lake metadata.
This means it can optimize your query by skipping the partitions that are not relevant for this query. Instead of having to read all 3 partitions, we can read only 1 and get the same result, just faster!
# Filter on partition columns will result in efficient partition pruning; non-matching partitions will be skipped.
> df.where(df["country"] == "Germany").show()
| | first_name | last_name | country | continent |
|---:|:-------------|:------------|:----------|------------:|
| 0 | Wolfgang | Manche | Germany | nan |
| 1 | Soraya | Jala | Germany | nan |
You can use the explain method to see how Daft is optimizing your query. Since we've already called collect on our DataFrame, it is already in memory. So below we copy the output of explain(show_all=True) before calling collect:
Running df.where(df["continent"] == "Asia").explain(True) returns:
(...)
== Optimized Logical Plan ==
* PythonScanOperator: DeltaLakeScanOperator(None)
| File schema = first_name#Utf8, last_name#Utf8, country#Utf8, continent#Utf8
| Partitioning keys = [PartitionField(country#Utf8)]
| Filter pushdown = col(continent) == lit("Asia")
| Output schema = first_name#Utf8, last_name#Utf8, country#Utf8, continent#Utf8
== Physical Plan ==
* TabularScan:
| Num Scan Tasks = 3
| Estimated Scan Bytes = 3045
| Clustering spec = { Num partitions = 3 }
Whereas running df.where(df["country"] == "Germany").explain(True) returns:
(...)
== Optimized Logical Plan ==
* PythonScanOperator: DeltaLakeScanOperator(None)
| File schema = first_name#Utf8, last_name#Utf8, country#Utf8, continent#Utf8
| Partitioning keys = [PartitionField(country#Utf8)]
| Partition Filter = col(country) == lit("Germany")
| Output schema = first_name#Utf8, last_name#Utf8, country#Utf8, continent#Utf8
== Physical Plan ==
* TabularScan:
| Num Scan Tasks = 1
| Estimated Scan Bytes = 1025
| Clustering spec = { Num partitions = 1 }
Running a query on a non-partitioned column like continent will require reading in all partitions, totalling 3045 bytes in this case.
Instead, running a query on a partitioned column (country in this case) means Daft only has to read only the relevant partition, saving us a whopping 2000+ bytes in this toy example :)
You can read High-Performance Querying on Massive Delta Lake Tables with Daft for an in-depth benchmarking of query optimization with Delta Lake and Daft.
Transform columns with Expressions
Daft provides a flexible Expressions API for defining computation that needs to happen over your columns.
For example, we can use daft.col() expressions together with the with_column method to create a new column full_name, joining the contents of the last_name column to the first_name column:
> df_full = df.with_column("full_name", daft.col('first_name') + ' ' + daft.col('last_name'))
> df_full.show()
| | first_name | last_name | country | continent | full_name |
|---:|:-------------|:------------|:----------|:------------|:----------------|
| 0 | Ernesto | Guevara | Argentina | NaN | Ernesto Guevara |
| 1 | Bruce | Lee | China | Asia | Bruce Lee |
| 2 | Jack | Ma | China | Asia | Jack Ma |
| 3 | Wolfgang | Manche | Germany | NaN | Wolfgang Manche |
| 4 | Soraya | Jala | Germany | NaN | Soraya Jala |
Multimodal Data Type Support
Daft has a rich multimodal type-system with support for Python objects, Images, URLs, Tensors and more.
Daft columns can contain any Python objects. For example, let's add a column containing a Python class Dog for some of the people in our dataset:
> import numpy as np
> class Dog:
> def __init__(self, name):
> self.name = name
> def bark(self):
> return f"{self.name}!"
> df_dogs = daft.from_pydict({
> 'full_name': ['Ernesto Guevara','Bruce Lee','Jack Ma','Wolfgang Manche','Soraya Jala'],
> "dogs": [Dog("ruffles"), Dog("shnoodles"), Dog("waffles"), Dog("doofus"), Dog("Fluffles")],
> })
> df_dogs.show()
| | full_name | dogs |
|---:|:----------------|:-------------------------------------|
| 0 | Ernesto Guevara | <__main__.Dog object at 0x1603d1c10> |
| 1 | Bruce Lee | <__main__.Dog object at 0x126ab9b90> |
| 2 | Jack Ma | <__main__.Dog object at 0x1603d27d0> |
| 3 | Wolfgang Manche | <__main__.Dog object at 0x1603d1cd0> |
| 4 | Soraya Jala | <__main__.Dog object at 0x1603d3f50> |
You can join this new dogs column to your existing DataFrame using the join method:
> df_family = df_full.join(df_dogs, on=["full_name"])
> df_family.show()
| | full_name | first_name | last_name | country | continent | dogs |
|---:|:----------------|:-------------|:------------|:----------|:------------|:-------------------------------------|
| 0 | Ernesto Guevara | Ernesto | Guevara | Argentina | NaN | <__main__.Dog object at 0x1603d1c10> |
| 1 | Bruce Lee | Bruce | Lee | China | Asia | <__main__.Dog object at 0x126ab9b90> |
| 2 | Jack Ma | Jack | Ma | China | Asia | <__main__.Dog object at 0x1603d27d0> |
| 3 | Wolfgang Manche | Wolfgang | Manche | Germany | NaN | <__main__.Dog object at 0x1603d1cd0> |
| 4 | Soraya Jala | Soraya | Jala | Germany | NaN | <__main__.Dog object at 0x1603d3f50> |
We can then use the apply method to apply a function to each instance of the Dog class:
> from daft import DataType
> df_family = df_family.with_column(
> "dogs_bark_name",
> df_family["dogs"].apply(lambda dog: dog.bark(), return_dtype=DataType.string()),
> )
> df_family.show()
| | first_name | last_name | country | continent | full_name | dogs | dogs_bark_name |
|---:|:-------------|:------------|:----------|:------------|:----------------|:-------------------------------------|:-----------------|
| 0 | Ernesto | Guevara | Argentina | NaN | Ernesto Guevara | <__main__.Dog object at 0x1603d1c10> | ruffles! |
| 1 | Bruce | Lee | China | Asia | Bruce Lee | <__main__.Dog object at 0x126ab9b90> | shnoodles! |
| 2 | Jack | Ma | China | Asia | Jack Ma | <__main__.Dog object at 0x1603d27d0> | waffles! |
| 3 | Wolfgang | Manche | Germany | NaN | Wolfgang Manche | <__main__.Dog object at 0x1603d1cd0> | doofus! |
| 4 | Soraya | Jala | Germany | NaN | Soraya Jala | <__main__.Dog object at 0x1603d3f50> | Fluffles! |
Daft DataFrames can also contain many other data types, like tensors, JSON, URLs and images. The Expressions API provides useful tools to work with these data types.
Take a look at the notebook in the delta-examples Github repository for a closer look at how Daft handles URLs, images and ML applications.
Transform multiple columns with UDFs
You can use User-Defined Functions (UDFs) to run functions over multiple rows or columns:
> from daft import udf
> @udf(return_dtype=DataType.string())
> def custom_bark(dog_series, owner_series):
> return [
> f"{dog.name} loves {owner_name}!"
> for dog, owner_name
> in zip(dog_series.to_pylist(), owner_series.to_pylist())
> ]
> df_family = df_family.with_column("custom_bark", custom_bark(df_family["dogs"], df_family["first_name"]))
> df_family.select("full_name", "dogs_bark_name", "custom_bark").show()
| | full_name | dogs_bark_name | custom_bark |
|---:|:----------------|:-----------------|:-----------------------|
| 0 | Ernesto Guevara | ruffles! | ruffles loves Ernesto! |
| 1 | Bruce Lee | shnoodles! | shnoodles loves Bruce! |
| 2 | Jack Ma | waffles! | waffles loves Jack! |
| 3 | Wolfgang Manche | doofus! | doofus loves Wolfgang! |
| 4 | Soraya Jala | Fluffles! | Fluffles loves Soraya! |
Daft supports workloads with many more data types than traditional DataFrame APIs.
By combining multimodal data support with the UDF functionality you can run ML workloads right within your DataFrame.
When should I use Daft DataFrames?
Daft DataFrames are designed for multimodal, distributed workloads.
You may want to consider using Daft if you're working with:
- Large datasets that don't fit into memory or would benefit from parallelization
- Multimodal data types, such as images, JSON, vector embeddings, and tensors
- ML workloads that would benefit from interactive computation within DataFrame (via UDFs)
Take a look at the Daft tutorials for in-depth examples of each use case.
Contribute to daft
Excited about Daft and want to contribute? Join them on Github 🚀
Like many technologies, Daft collects some non-identifiable telemetry to improve the product. This is stricly non-identifiable metadata. You can disable telemetry by setting the following environment variable: DAFT_ANALYTICS_ENABLED=0. Read more in the Daft documentation.