Skip to main content

Yummy-Delta

This repository contains Yummy Delta Lake api written in Rust. The api use delta-rs implementation.

Instalation

pip3 install yummy-delta

Quick start

import yummy_delta
# yummy_delta.run({config_path}, {host}, {port}, {log_level})
yummy_delta.run("config.yaml", "0.0.0.0", 8080, "error")

In the configuration file we can define multiple stores using supported backends.

config.yaml

stores:
  - name: local
    path: "/tmp/delta-test-1/"
  - name: az
    path: "az://delta-test-1/"

Delta backends

Local file system

/tmp/delta-test-1

AWS S3 / S3

s3://<bucket>/<path>
s3a://<bucket>/<path>

Environment variables: AWS_REGION="us-east-1", AWS_ACCESS_KEY_ID="deltalake", AWS_SECRET_ACCESS_KEY="weloverust", AWS_ENDPOINT_URL=endpoint_url,

Azure Blob Storage / Azure Datalake Storage Gen2

az://<container>/<path>
adl://<container>/<path>
abfs://<container>/<path>

Environment variables: AZURE_STORAGE_ACCOUNT_NAME="devstoreaccount1", AZURE_STORAGE_ACCOUNT_KEY="****", AZURE_STORAGE_USE_EMULATOR="false", AZURE_STORAGE_USE_HTTP="true",

Google cloud storage

gs://<bucket>/<path>

Rest API

Stores list

Request:

curl -X GET "http://localhost:8080/api/1.0/delta" \
-H "Content-Type: application/json"

Response:

{'stores': [{'path': '/tmp/delta-test-1/', 'store': 'local'},
            {'path': 'az://delta-test-1/', 'store': 'az'}]}

Tables list

Request:

curl -X GET "http://localhost:8080/api/1.0/delta/{store_name}" \
-H "Content-Type: application/json"

Response:

{'path': 'az://delta-test-1/',
 'store': 'az',
 'tables': ['test_delta_1', 'test_delta_2', 'test_delta_3', 'test_delta_4']}

Table details

Request:

curl -X GET "http://localhost:8080/api/1.0/delta/{store_name}/{table_name}/details" \
-H "Content-Type: application/json"

Response:

{'path': 'az://delta-test-1/test_delta_4',
 'schema': {'fields': [{'metadata': {},
                        'name': 'col1',
                        'nullable': false,
                        'type': 'integer'},
                       {'metadata': {},
                        'name': 'col2',
                        'nullable': false,
                        'type': 'string'}],
            'type': 'struct'},
 'store': 'az',
 'table': 'test_delta_4',
 'version': 100}

Create delta table

Request:

curl -X POST "http://localhost:8080/api/1.0/delta/{store_name}" \
-H "Content-Type: application/json" \
-d '{
    "table": "test_delta_4",
    "schema": [
        {
            "name": "col1",
            "type": "integer",
            "nullable": false
        },
        {
            "name": "col2",
            "type": "string",
            "nullable": false
        }
    ],
    "partition_columns": ["col2"],
    "comment": "Hello from delta"
}'

Response:

{ "table": "test_delta_4" }

Append data

Request:

curl -X POST "http://localhost:8080/api/1.0/delta/{store_name}/{table_name}" \
-H "Content-Type: application/json" \
-d '
{"record_batch_dict": {"col1": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11],
                       "col2": ["A", "B", "A", "B", "A", "A", "A", "B", "B", "A", "A"]}}'

Response:

{'table': 'test_delta_4', 'version': 101}

OR:

Request:

curl -X POST "http://localhost:8080/api/1.0/delta/{store_name}/{table_name}" \
-H "Content-Type: application/json" \
-d '
{"record_batch_list": [
    {"col1": 1, "col2": "A"}, 
    {"col1": 2, "col2": "B"}, 
    {"col1": 3, "col2": "A"}, 
    {"col1": 4, "col2": "B"}, 
    {"col1": 5, "col2": "A"}, 
    {"col1": 6, "col2": "A"}, 
    {"col1": 7, "col2": "A"}, 
    {"col1": 8, "col2": "B"}, 
    {"col1": 9, "col2": "B"}, 
    {"col1": 10, "col2": "A"}, 
    {"col1": 11, "col2": "A"}
    ]
}

Response:

{'table': 'test_delta_4', 'version': 101}

Overwrite data

Request:

curl -X PUT "http://localhost:8080/api/1.0/delta/{store_name}/{table_name}" \
-H "Content-Type: application/json" \
-d '
{"record_batch_dict": {"col1": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11],
                       "col2": ["A", "B", "A", "B", "A", "A", "A", "B", "B", "A", "A"]}}'

Response:

{'table': 'test_delta_4', 'version': 101}

OR:

Request:

curl -X PUT "http://localhost:8080/api/1.0/delta/{store_name}/{table_name}" \
-H "Content-Type: application/json" \
-d '
{"record_batch_list": [
    {"col1": 1, "col2": "A"}, 
    {"col1": 2, "col2": "B"}, 
    {"col1": 3, "col2": "A"}, 
    {"col1": 4, "col2": "B"}, 
    {"col1": 5, "col2": "A"}, 
    {"col1": 6, "col2": "A"}, 
    {"col1": 7, "col2": "A"}, 
    {"col1": 8, "col2": "B"}, 
    {"col1": 9, "col2": "B"}, 
    {"col1": 10, "col2": "A"}, 
    {"col1": 11, "col2": "A"}
    ]
}

Response:

{'table': 'test_delta_4', 'version': 101}

Optimize

Request:

curl -X POST "http://localhost:8080/api/1.0/delta/{store_name}/{table_name}/optimize" \
-H "Content-Type: application/json" \
-d '{
    "target_size": 2000000,
    "filters": [
        {
            "column": "col2",
            "operator": "=",
            "value": "B"
        }
    ]
}'

Response:

{'metrics': {'filesAdded': {'avg': 524.5,
                            'max': 541,
                            'min': 508,
                            'totalFiles': 2,
                            'totalSize': 1049},
             'filesRemoved': {'avg': 506.75,
                              'max': 538,
                              'min': 499,
                              'totalFiles': 36,
                              'totalSize': 18243},
             'numBatches': 1,
             'numFilesAdded': 2,
             'numFilesRemoved': 36,
             'partitionsOptimized': 2,
             'preserveInsertionOrder': true,
             'totalConsideredFiles': 36,
             'totalFilesSkipped': 0}}

Vacuum

Request:

curl -X POST "http://localhost:8080/api/1.0/delta/{store_name}/{table_name}/vacuum" \
-H "Content-Type: application/json" \
-d '{
    "retention_period_seconds": 0,
    "enforce_retention_duration": false,
    "dry_run": false
}'

Response:

{'dry_run': False,
 'files_deleted': ['col2=A/part-00000-09fef82b-9208-41be-8769-30c442b36346-c000.snappy.parquet',
                   'col2=A/part-00000-16779a2f-d2f4-450c-b815-3b8fca850402-c000.snappy.parquet',
                   'col2=A/part-00000-26e01b49-a432-4d06-816d-9a2048c0194e-c000.snappy.parquet',
                   'col2=A/part-00000-3ff1f905-87ae-4b73-a3ae-3cd1056171a0-c000.snappy.parquet',
                   'col2=A/part-00000-50dbcfea-22d6-4ef3-bacf-f1371b9186ad-c000.snappy.parquet',
                   'col2=A/part-00000-511d3ceb-0959-4024-9bbe-d049bb4fb4c6-c000.snappy.parquet',
                   'col2=A/part-00000-56e5484f-1722-45eb-84bb-1475b67e9f63-c000.snappy.parquet',
                   'col2=B/part-00000-b15a1cd7-9a15-4ba6-8ba8-e3ca63737f81-c000.snappy.parquet',
                   'col2=B/part-00000-d4fadc68-46b5-4f1f-a378-e9f3dba842f0-c000.snappy.parquet',
                   'col2=B/part-00000-d6e823d1-433e-4798-a641-1e2783c8c1eb-c000.snappy.parquet',
                   'col2=B/part-00000-ded9a5b7-4754-47b7-9e2c-132263cea52e-c000.snappy.parquet',
                   'col2=B/part-00000-fb89f02c-8ea3-47d9-8794-182d8602687c-c000.snappy.parquet']}

Metadata

Release files for yummy-delta 0.0.9

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Built distribution (wheel)

Table of built distributions (wheels) for yummy-delta 0.0.9
File Interpreter ABI Platform Actions
yummy_delta-0.0.9-cp39-cp39-manylinux_2_31_x86_64.whl CPython 3.9 CPython 3.9 Linux glibc 2.31+ x86-64 Details for yummy_delta-0.0.9-cp39-cp39-manylinux_2_31_x86_64.whl

Release files / yummy_delta-0.0.9-cp39-cp39-manylinux_2_31_x86_64.whl

Download URL yummy_delta-0.0.9-cp39-cp39-manylinux_2_31_x86_64.whl
Size 22.7 MB
Tags CPython 3.9 Linux glibc 2.31+ x86-64
SHA-256 checksum
How to use checksums
b9ea6c58308d78b6f12f00cc9ea405ab23961385a922e7a08e46b048f732b826
BLAKE2b-256 checksum
How to use checksums
c57bd9190d8eb51a4a4d0cb94dc4c9632361f5e44cc4c66e3fb75fe8d886e812
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/4.0.2 CPython/3.9.2

Security for yummy-delta 0.0.9

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page