When will Elasticsearch SQL be released?

I am looking for the Elasticsearch SQL that was introduced during last Elastic{ON}17.
Is it already available ? It has been a while since the show demo.
I see that there are community alternatives but what I am looking for is a jdbc connector.

1 Like

Hey @christophe_spielmann. You haven't missed it yet! It's still not released. We don't have an ETA we can share publicly, but we're working hard to move it forward as quickly as possible.

A few questions though would help understand what you need:

  1. Do you intend to hook JDBC into a well-known app (e.g. a particular BI tool) or is it for a home-grown thing you're building up? If home-grown, which client language(s) are you using?
  2. What types of SQL statements and functions do you need? The initial target is for query only (e.g. SELECT) rather than data/cluster modifications (e.g. INSERT) and we're interested to know if there are specific functions/predicates that folks need, since they each need to be implemented one-by-one.

Any other insights into what you want to do or how you expect to use it would help us understand and build a better product!

thanks Shane for the answer.

  1. Yes I would like to hook es-jdbc into jasper report sothat I can report ES data through usual static reports.
    The point is that we enable a datamart for each of our customer sothat they can use the API to get their data and do whatever they want. No need anymore to do specfic report/data push for each of our customer. We provide as well a Kibana (multitenancy) for each of them sothat they can create their dashboard. But for some customers it is either too big a shift in their mind or they don't have the IT maturity to do that. Hence as fall back we would be able to push them reports as usual. The report feature currently into Xpack kibana is rather limited and does not fulfill that requirement. And moreover kibana multitenancy pattern makes Xpack reporting fail.
  2. Only BI is foreseen hence select only is required. Regarding the functions/predicates required, i think only basic ones are required ( listing, grouping, basic aggregation). But I didn't do a rationale about that yet.

I am really eager to get this jdbc for ES select only.

Note: I have to tell that I didn't do a tour about other integration solution of ES data into BI solutions. I was so happy when I saw you demo video: I can keep my current BI solution (jasper) during the transition stage of all our customers to ES rest API.

Christophe,

If you are interested in connecting Jasper to ES, commercial JDBC drivers are already available: http://www.cdata.com/drivers/elasticsearch/jdbc/. The CData JDBC Driver for ES includes a rich SQL92 engine with support for query normalization & complex queries, including JOIN, GROUP BY, ORDER BY, and all filters (including filters with formulas). CRUD operations are even supported.

More details about using the JDBC driver in Jasper Studio are available here: http://www.cdata.com/kb/tech/elasticsearch-jdbc-jaspersoft-studio.rst

This is quite interesting. If ES JDBC driver should not update/write, but other ES API accesses are still allowed to write while ES JDBC driver is active, how can SQL transaction isolation levels be implemented? E.g. read uncommitted, read committed? Will always dirty reads be allowed? And phantom reads?

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.