I am using ELK stack to perform some analysis on my data. My index pattern has 2 types of data
The ID of the product clicked
The ID of the product sold
I would like to first find all the IDs sold out and based on that find the number of times that ID was clicked.
Example-
1, clicked
1, sold
2, clicked
2, clicked
2, sold
3. clicked
So in this data-> ID1 and ID2 were sold. And ID1 was clicked once and ID2 was clicked twice.
You can achieve this by using filter aggregations in a visualization.
Either in Lens or even in normal Bar Chart.
You do a terms Aggregation on the product ID, with a count metric. Then you can split that aggregation with a filter one where the first filter is "action: clicked" and the second one is "action: sold".
This will give me 2 Aggregations one for 'sold' and the other one for 'clicked'.
I would like to get the Number of clicks for all the Sold Items only. When I apply a filter on 'sold', I get only sold rows. Now how can I get the data for the click? Its similar to sub-queries in SQL.
I checked many posts, and all say that Sub-Queries in ES is not possible. Moreover, if I start using "Canvas" and "SQL", it does not support all the Sub-Queries.
As of now, it seems like I will have to store all the data on a database (MySql or MongoDB) and use Grafana for Visualization.....
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.