calcite-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Julian Hyde <jul...@hydromatic.net>
Subject Re: Guidance for writing a Calcite adapter for a deep hierarchical dataset
Date Wed, 08 Oct 2014 19:41:33 GMT
On Oct 8, 2014, at 12:05 PM, Hartman, Trevor <thartman@ebay.com> wrote:

> Ah, aren't MongoDB and JSON both examples of hierarchical datasets?

Yes. You can handle these if you work in “type-less” mode. Each column is a map, which
contains (property, value) pairs, and each value might be regular a value, an array, or another
map. It is very close to the JSON model.

That’s what I’d suggest for your case.

A useful addition would be something like Oracle’s JSON_VALUE function [1], which allows
you to use something like an XPath expression to burrow deep into a nested document.

And by the way, you asked about schemas. Some people use it to mean “the set of columns
and their types in a table”, but we use the term “row type” for that. A calcite schema,
just as in JDBC [2], is a namespace that contains tables. (Similar to a catalog and a database.)
I think you are interested in what we call a row type.

Julian

[1] http://docs.oracle.com/database/121/ADXDB/json.htm#ADXDB6263
[2] http://docs.oracle.com/javase/7/docs/api/java/sql/Connection.html#getSchema() 
Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message