Joining queries

[amit.patra (Amit Kumar Patra)]

Hi,

I know that performing full SQL-style joins in a distributed system like Elasticsearch is prohibitively expensive.

So I am giving an example of from RDBMS concept. Can you please suggest that what will be the efficient data model to avoid join in elasticsearch.

A table name: country_code and the structure is

country_name varchar(70) DEFAULT NULL,
country_code varchar(8) DEFAULT NULL,
image blob DEFAULT NULL,
country_phcode varchar(8) DEFAULT NULL

Another table name : mms_cri and the structure is

id varchar(100) NOT NULL,
direction varchar(1) DEFAULT NULL,
country_code varchar(60) DEFAULT NULL
SATELITE_NAME varchar(60) DEFAULT NULL,
TERMINAL_TYPE varchar(40) DEFAULT NULL,
T2TID varchar(1024) DEFAULT NULL,
T2THOP varchar(10) DEFAULT NULL,

In RDBMS , We join above two table using country_code.

In the above case scenario , How efficient data modelling can be done ? Can you explain with an example.

Thanks,

[Christian_Dahlqvist]

Storing large blobs in Elasticsearch is not recommended, so assuming the image attribute is instead a path to there the blob is located, I would recommend simply denormalizing the data.