API Conventions in Elasticsearch

In the web, an API is a set of function calls to access software components in a particular application. For example, Facebook API allows the developers to access the data (like DOB or status update) or other functionalities from Facebook to create applications. Elasticsearch offers the REST API and it also uses some conventions that are discussed below in this chapter.

The Elasticsearch REST APIs are exposed through the JSON over HTTP . It uses the following conventions that are listed below in this chapter. These conventions can be applied throughout the REST API.

Multiple Indices

Date Math Support in Index Name

URL based Access Control

Common Options

Multiple Indices

In API, most of the operations like searching are for one or more indices. This helps the user to perform various operations (like searching) in multiple places through the entire API . The user can search all the available data by executing a search query for once.

For these queries, some notations are used, as given below:

Comma Separated Notation (,)

Wildcard Notation (*, +, -)

URL Query String Parameter

allow_no_indices

_all keywords for all indices

These different notations are used to perform operations on multiple indices.

Date Math Support in Index Name

This API convention allows the user to search for a range of time-series indices. This type of search limits the range of number of indices that are being searched instead of searching all your time-series indices. This will reduce the load on a cluster and improve the execution performance.

Almost all APIs support date math that contains an index

Elasticsearch provides this convention to search the indices according to date and time.

For which mathematical operation is used where we need to specify the date and time in a specific format to find the data for a particular date.

Expression: <static_name {date_math_expr {date_format|time_zone}}>

static_name

date_math_expr

date_format

time_zone

For example: <accountdetail - {now-2d { YYYY-MM-dd|utc }}>

static_name is a part of expression that cannot be changed. It remains same in each date math index.

date_math_expr contains the format of date (like dd-MM-YYYY) and time to be written in index. The default value is YYYY-MM-dd, which is changable.

Let's suppose today's date is 18 ^th May 2020, then it will return the account detail of 2020-05-16.

Expression Result <accountdetail - {now-d}> accountdetails - 2020-05-18 <accountdetail - {now-2M}> accountdetails - 2020-03-19 <accountdetail - {now- {YYYY-MM}}> accountdetails - 2020-05

URL-based Access Control

Many users use proxy with this convention for secure access to Elasticsearch indices.

Elasticsearch provides several API that allows the user to specify the individual request in request body such as multi_get, multi_search, and bulk requests, etc. The users have the choice to specify the index in URL , which makes URL-based Access Control more challenging.

Add the following setting in elasticsearch.yml file to disable the default action:

rest.action.multi_allow_explicit_index = false;

By default, it is set to true .

Common Options

Elasticsearch provides the following common options are:

Pretty Result

Human Readable Output

Date Math

Flat Settings

No Value

Time Unit

Parameter

Fuzziness

Distance Unit

Byte Size Units

Unit-less quantities

Enabling Stack Tracks

Response Filtering

Request Body in Query Setting

Now, we will discuss some common options in details:

Pretty Results

Append the following URL query parameter (pretty = true;) to get the result in a well-formatted JSON object.

pretty = true;

Human Readable Output

Human Readable Output option is used to change the statistical response in either computer-readable form or human-readable form.

If set human = false , it converts the statistical response to a computer-readable form.

If set human = true , it converts the statistical response to a human-readable form.

The default value of it is FALSE .

For Example -

If human = true, then distance_kilometer = 50KM If human = false, then distance_meter = 50000 Next Topic Elasticsearch vs Solr