kendo-elasticsearch

A Kendo DataSource extension so you can load data into your Kendo UI Grid from an ElasticSearch index.

It supports filtering, searching, sorting, grouping and aggregating in ElasticSearch for date, number (double only) and string fields.

Install

bower install kendo-elasticsearch-datasource

Demos

To run the demos on your computer you will need a local instance of ElasticSearch with CORS enabled, a clone of this repository and nodejs and bower.

If you use docker-compose this command will create an Elasticsearch instance:

docker-compose up -d

If you use docker this command will create an ElasticSearch instance :

docker run -d --name elasticsearch1.7 -p 9200:9200 elasticsearch:1.7-alpine -Des.http.cors.enabled=true -Des.http.cors.allow-origin=*

The dataset is constituted of 2 simple mappings: "organization" and "person". The documents are generated randomly using json-schema-faker. The persons have an organization as parent. Each textual field is indexed using the standard analyzer and has additional subfields more suitable for regexp filtering and aggregations on exact values. Have a look at the index definition to see how this is done.

Clone the project and create the datasets:

git clone https://github.com/MGDIS/kendo-elasticsearch.git
cd kendo-elasticsearch
npm install
bower install
node demos/create-datasets.js

Afterwards you just have to open the HTML files from the demos folder in your browser.

Basic

This example queries the "person" mapping. It has paging, sorting and filtering. Filtering of text fields is based on a "lowercase" subfield, except for the additional "search" operator which behaves as a classical keywords search.

See the source code.

Aggregate

This example illustrates using server side aggregations to work on a number field and the cardinality of a text field. Aggregations are dependent on the filters, but not on the pagination.

See the source code.

Groups and group aggregations

This example illustrates using server side bucket aggregations to group lines based on the values of a field. It also uses metric aggregations on the buckets.

Please note that it is better to use this functionality on fields with a moderate cardinality. This is because all buckets will be requested, not only the ones needed to display the current page.

See the source code.

Joined multi values

This example illustrates using a custom separator to join the multiple values of a field inside a single cell.

The separator can use HTML if the 'encoded' attribute of the column is set to 'false'.

See the source code.

Split multi values

This example illustrates splitting the lines of data based on the multiple values of a field.

Please note that the page size and filters can seem to be badly interpreted. This is because the actual lines of data fetched from the server and the lines displayed to the user are 2 separate things.

This is probably not a very useful feature for user interactions and visualization, but it can be handy for exporting the dataset.

See the source code.

Dates

This example illustrates working with date fields and grouping items based on a server-side date histogram.

Please note that kendo grid does not support grouping multiple times on the same field. At least it does not allow us to define a different group header each time. This is why multiple fields are defined in the model that actually redirect to the same field in ElasticSearch.

See the source code.

Nested

This example illustrates displaying information from an array of nested objects and arrays into a second level of nesting (addresses and addresses.telephones). The lines of data are duplicated much like in the "Split multi values" demo, so the page size is not visually respected. But the inner hits functionality of ElasticSearch is used so that filtering works well.

Note that due to this issue, it is necessary to use top level inner_hits syntax.

See the source code.

Parent

This example illustrates displaying information from a parent document.

Please note that on ElasticSearch pre 2.0 the inner hits functionality has a bug that prevents us from using a search URL that includes the type. See the related issue.

Also sorting on the parent is not possible yet, nor aggregating.

See the source code.

Children

This example illustrates displaying information from children documents.

Please note that on ElasticSearch pre 2.0 the inner hits functionality has a bug that prevents us from using a search URL that includes the type. See the related issue.

Also sorting on the child is not possible yet, or only only as a secondary sort inside the children of a same parent.

See the source code.

Mapping

This example illustrates passing a ElasticSearch mapping instead of kendo fields definitions.

The datasource will automatically transform one into the other.

See the source code.

Exists and missing filters

This example illustrates costumizing kendo's filters to support the 'exists' and 'missing' elasticsearch filters.

TODO:

• Add a note about kendo grid licence, web and pro packs.
• Add a note about ElasticSearch versions compatibility.
• Add notes about kendo/ES functionalities mapping and the rational behind it
• Combine with some external filter or query
• Move these todos into actual issues 😃