Can we expose REST API on detla tables?

[delta-io]

Can we expose REST API on detla tables?

[tdas]

We dont have any support for that right now. The only API we have write now is Spark DataFrame Reader Writer APIs, and we will have SQL support that will be executable through Spark SQL.

It's important to understand that Delta Lake is a data layout format. It does not run a service or a process, so the question of any API endpoint does not arise.

[tdas]

I am closing this issue. Please reopen it if you have any follow up questions.

[ReubenTheDS]

@tdas A year later, but I have a question in the same "zone" as the OP. I understand "that Delta Lake is a data layout format" (quoted above). Is there now a standard for "exposing" the large Delta Lake tables(we intend using HDFS for storage) - reason we'd like to run visualisations on Delta Lake tables created by Spark scripts, rather than:
querying HDFS -> porting Delta Lake table snippet to Spark Parquet -> writing temp parquet data-> visualise using a 3rd party platform -> delete temp Parquet data

Or is there another way you suggest we go about this? Thanks!

[rjurney]

This would be really helpful and really powerful. We could store schemas in one place - Delta Tables - rather than more than one place. We just need to be able to fetch the schema from a Delta table to do this :(

[zsxwing]

@rjurney Delta Lake is just a table format, similar to Parquet. It doesn't have any service. If you would like to read Delta tables through REST APIs, you can try Delta Sharing.