dask-elk

Use dask to fetch data from Elasticsearch in parallel by sending the request to each shard separatelly.

Introduction
Usage

Introduction

The library tries to imitate the functionality of the ES Hadoop plugin for spark. dask-elk performs a parallel read across all the target indices shards. In order to achieve that it uses Elasticsearch scrolling mechanism.

Usage

To use the library and read from an index:

from dask_elk.client import DaskElasticClient

# First create a client
client = DaskElasticClient() # localhost Elasticsearch

index = 'my-index'
df = client.read(index=index, doc_type='_doc')

You can even pass a query to push down to elasticsearch, so that any filtering can be done on the Elasticsearch side. Because dask-elk uses scroll mechanism aggregations are not supported

from dask_elk.client import DaskElasticClient

# First create a client
client = DaskElasticClient() # localhost Elasticsearch
query = {
    "query" : {
        "term" : { "user" : "kimchy" }
    }
}
index = 'my-index'
df = client.read(query=query, index=index, doc_type='_doc')

Read documentation here

vincentsun870530/dask-elk

dask-elk

Table of Contents

Introduction

Usage