Flashback/Look-ahead – SQL: The Kingmaker of NoSQL Databases?

Over 1 ½ years ago I suggested that SQL will turn out to be an important factor for the success of NoSQL databases (https://realprogrammer.wordpress.com/2014/03/04/sql-the-kingmaker-of-nosql-databases/). Did that happen?

NoSQL / OldSQL / NewSQL: SQL Language Support on the Rise

When surveying a subset of the various database systems it becomes apparent that the support for relational SQL as well as JSON-supporting SQL is increasing. Right now this does not necessarily mean that there is an agreement on the particular language syntax, its semantic interpretation or execution style (declarative vs. functional vs. procedural). However, there is an (implicit) agreement that a (declarative) query language is necessary or advantageous, that SQL is a guiding principle and that JSON needs to be supported to various extent. A (not necessarily complete) list of different systems follows that have a (JSON-supporting) SQL language implemented in one form or another:

From the above selection it becomes clear that a lot of NoSQL, OldSQL, and NewSQL systems invest in SQL as well as JSON-supporting SQL and that the number of systems that support (JSON-supporting) SQL is definitely increasing.

Given that implementing the efficient execution of SQL is a significant engineering effort (and investment), JSON-supporting SQL is obviously seen as an important feature.

As an aside, an intriguing question is: why do database systems that fall into the category of “NoSQL” databases start providing SQL? It appears that the “No” in “NoSQL” is being re-interpreted from “no-SQL” to “no-Relational Model” with JSON-supporting SQL support. One possible answer to the question is the increasing complexity of queries that are being issued. While simple interfaces are good to get going initially, over time, as applications become more involved, complex queries reflect the increasing domain complexity and that is going to be reflected in the query language as well (for a taste, see here: https://realprogrammer.wordpress.com/2014/10/27/oracle-12c-sql-for-json-part-2-basic-queries/ – toward the end of the blog). Another (additional) possible answer is that existing tools (e.g. reporting tools) that are JDBC compliant and can process relational data should be able to connect also.

JSON-supporting SQL Language: Discussions are Ongoing

Discussions about the benefit of JSON-supporting SQL in context of NoSQL databases are coming up repeatedly (e.g., LinkedIn group discussions: https://www.linkedin.com/groups/2085042 (NoSQL Group)). Interesting enough, most of the time, the discussion is not separated or structured by syntax vs. semantics vs. functionality. SQL is developed in context of the relational model and that does not preclude complex data structures like JSON, the opposite is true. Object / relational functionality as well as NF2 work deals with complex structures, including arrays and maps. So it would be prudent, in my mind, to have a separation in the discussion of

  • Data structure support (aka, JSON)
  • Query functionality
  • Query language syntax

I expect the discussion of JSON-supporting SQL to lead towards a declarative language with SQL syntax, based on the relational model, paired with complex data types (e.g. JSON) and a schema-per-object or schema-per-row semantics.

The interesting aspect on the SQL syntax and interpretation is that systems that also support the relational model extend those to support JSON, and these systems will extend SQL, not replace it (as can be seen already).

SQL Support for NoSQL Databases: Best Approach Forward?

In my mind the best approach forward would be standing on the shoulder of giants: take SQL and extend it to support JSON in a very careful but very expressive way, learning from the object / relational work as well as the NF2 work that took place a long time back. Stay true to the learnings of the benefits of a declarative language and the underlying possibility of optimization, be it functional or non-functional. In short, add the domain type JSON and cater for a schema-per-document model.

If JSON support is one of the common functional denominators of most of the databases then the distinction is not really achieved by variations of (JSON-supporting) SQL, but by variations of the non-functional support like scale, throughput, low latency, safety, reliability, etc.

I believe that JSON-supporting SQL will be converging, while the non-functional features will set the various systems further and further apart in order to provide the designed-for sweet spot and establish their competitive advantage.


Clearly, the jury is still out, however, compared to 1 ½ years ago, whispers are emerging that JSON-supporting SQL is becoming an important feature of NoSQL, OldSQL and NewSQL databases.

Clearly, the pendulum swings toward the SQL language grammar and functionality that supports JSON. And that is a good thing in my opinion for the various obvious reasons.

Clearly, discussions keep coming up about JSON-supporting SQL, the various styles and language proposals. This is another reason to believe that SQL support in NoSQL databases and in context of JSON is important.

Let’s check in again into this topic next year some time. It is always interesting to follow such a hotly debated topic and establish checkpoints along the way..



The views expressed on this blog are my own and do not necessarily reflect the views of Oracle.


Oracle 12c – SQL for JSON (Part 3): Basic Joins

Having JSON support in a relational database means that the join operator is available not only for joining relational data or JSON data, but also for the mixed case: joining JSON and relational data. This opens up a whole new world of data modeling and query execution.

Running Example

This running example creates three tables, “demo”, “city” and “city_rel” and a sample data set in each table. The tables “city” and “city_rel” contain the same data set, once in JSON format, and once in relational format.

  id NUMBER,
  person CLOB 
    CONSTRAINT person_ensure_json 
( 1, '{ "name": "Bob", "city": "SF"}' );
( 2, '{ "name": "Jake", "city": "PA"}' );
( 3, '{ "name": "Alice", "city": "NYC"}' );
( 4, '{ "name": "Jenn",  "city": {"name": "Tokyo"}}' );
( 5, '{ "name": "Jenn",  "city": ["Tokyo"]}' );
( 6, '{ "name": "Jenn",  "city": 66}' );
  id NUMBER,
  city CLOB 
    CONSTRAINT city_ensure_json 
( 101, '{"city": "SF", "state": "CA", 
  "country": "US"}' );
( 102, '{"city": "PA", "state": "CA", 
  "country": "US"}' );
( 103, '{"city": "NYC", "state": "NY", 
  "country": "US"}' );
( 104, '{"city": {"name": "Tokyo"}, "state": null, 
  "country": "Japan"}' );
( 105, '{"city": ["Tokyo"], "state": null, 
  "country": "Japan"}' );
( 106, '{"city": 66, "state": null, 
  "country": "World"}' );
DROP TABLE city_rel;
  id      NUMBER,
  city    VARCHAR(255),
  state   VARCHAR(255),
  country VARCHAR(255));
( 1001, 'SF', 'CA', 'US' );
( 1002, 'PA', 'CA', 'US' );
( 1003, 'NYC', 'NY', 'US' );
( 1004, '{"name": "Tokyo"}', NULL, 'World' );
( 1005, '["Tokyo"]', NULL, 'World' );
( 1006, '66', NULL, 'World' );


The following SQL statement is a simple join between JSON structures on the property “city”:

FROM demo d, city c
WHERE d.person.city = c.city.city;

This SQL statement projects in addition to joining JSON structures:

SELECT d.person, c.city
FROM demo d, city c
WHERE d.person.city = c.city.city;

The following SQL statement extends the projection:

FROM demo d, city c
WHERE d.person.city = c.city.city;

JSON – Relational Join

 This SQL statement shows the join between JSON and relational data, combined with a projection:

FROM demo d, city_rel c_r
WHERE d.person.city = c_r.city;

Significance of Pure and Mixed JSON Joins

As shown, the join operator is applied easily within JSON tables and across JSON and relational tables. When using Oracle 12c there is no restriction anymore when it comes to the join operator in conjunction of JSON documents.

Pure JSON joins are possible in context of Oracle 12c. This means that developers have a choice to model all data in a pure document form (trying to avoid the need for joins by creating sub-collections – which is almost impossible without denormalization), or to consciously model documents in such a way that the document nature is applied where applicable without having to necessarily de-normalize as the join operator is available.

The mixed case between JSON and relational tables goes a lot further as now data can be modeled according to its nature (not all data is exclusively document-oriented or relational) and its access path requirements without compromising either way.

In addition, the mixed case supports the situation where data is already present in the database in relational form and new data is added in JSON form. This means that even if data is available in relational form, additional data does not have to be in relational form, and the most appropriate representation can be chosen (and no separate document-oriented database has to be deployed, btw).



The views expressed on this blog are my own and do not necessarily reflect the views of Oracle.

NoSQL: No-Tation

This is too good to pass as there is a ton of truth in this simple presentation below. It is cited from http://java.dzone.com/articles/history-databases-”no-tation” without modification:


What Flavor of SQL?

The question now is: what flavor of SQL is the SQL in “No, SQL!”?

One alternative is to start from scratch and invent a SQL grammar and semantics that is not based on the existing SQL standard and SQL implementations. That would not be my choice at all, for the obvious reasons.

Another alternative is to treat JSON ‘just’ as a data structure and embed it into existing SQL. This would be roughly the equivalent of the object/relational model (~NF2) with the variation that there is no global schema and some non-standard SQL types.

The (in my opinion) better alternative is to look at JSON as more than just a data structure and combine JSON with the schema-per-document paradigm as the underlying philosophy.

  • This would be a real game changer as it would cater to the dynamic changing JSON data structures as needed in many projects.
  • As a result, the schema-per-document model would be integrated with the relational model so that a single database management system can support several data models and their paradigms concurrently.

Extending SQL Semantics

Supporting a schema-per-document philosophy requires extending the SQL semantics and its operators to be able to deal with the fact that different JSON structures might have a different schema, even if they are in the same collection or relational table.

Operators like typeOf() come to mind that can test if a path into a JSON structure refers to a specific JSON type or JSON literal name. Another operator is exists() to test if a path into a JSON structure exists. Or array operators that can determine if an array contains a specific JSON structure as an element. containsPath() is an operator that checks if a JSON structure has the path given.

Many more operators can be envisioned to make managing and querying JSON structures in context of SQL expressive and useful.



The views expressed on this blog are my own and do not necessarily reflect the views of Oracle.


Document-oriented NoSQL Databases: Oracle NoSQL’s Shift in March 2014

I believe Oracle NoSQL extended and expanded from a key/value store to a document-oriented NoSQL database in March 2014. If you concur, then this was a major shift in the NoSQL arena of document-oriented databases.

Document-Oriented NoSQL Databases

In a narrow interpretation, document-oriented NoSQL databases use JSON or BSON as the query interface data representation. In a wider interpretation, any scalar data type and any composite data type (like maps or array) is available at the query interface without those necessarily complying to the JSON encoding or interpretation.

Some of the known document-oriented database are MongoDB, CouchDB and RethinkDB, amongst many others. The site http://db-engines.com/en/ranking/document+store has a categorization and ranking readily available.

In addition to storing documents (aka, complex data structures with scalar, array and map data types), another important criteria is the ability to query based on the document content by e.g. selecting documents where specific properties have specific values or comply to specific complex predicates (in the general case).

Key/Value Stores

Key/value stores (k/v stores) are different from document-oriented databases. Key/value stores persist a value for a given key. Retrieval is based on the key and values cannot be used in query predicates. Some k/v stores have structured keys or composite keys that support querying a range or set of values.

Oracle NoSQL, for example, is supporting structured and composite keys (http://docs.oracle.com/cd/NOSQL/html/GettingStartedGuide/singleget.html, http://docs.oracle.com/cd/NOSQL/html/GettingStartedGuide/multiget.html).

Aerospike is another example of a k/v store (https://docs.aerospike.com/display/V3/Key-Value+Store+Guide).

K/V Stores: Major Shift

Oracle NoSQL extended its data model from a pure k/v data model to a document-oriented data model in March 2014.

The available simple data types are (cited from http://docs.oracle.com/cd/NOSQL/html/GettingStartedGuideTables/tablesapi.html)

  • Double
  • Float
  • Integer
  • Long
  • Java byte array
  • Java String

And the available complex data types are (cited from http://docs.oracle.com/cd/NOSQL/html/GettingStartedGuideTables/tablesapi.html)

  • Array (An array of values, all of the same type)
  • Enum (An enumeration, represented as an array of strings)
  • Fixed Binary (A fixed-sized binary type used to handle binary data where each record is the same size. It uses less storage than an unrestricted binary field, which requires the length to be stored with the data)
  • Map (An unordered map type where all entries are constrained by a single type)
  • Records

In addition, indexes can be created on the ‘value’ part, aka, the documents: http://docs.oracle.com/cd/NOSQL/html/GettingStartedGuideTables/createindex.html. And based on indexes, the database can be queried without having to specify a key: http://docs.oracle.com/cd/NOSQL/html/GettingStartedGuideTables/indexread.html.

The latter functionality extends the characteristics of the Oracle NoSQL database from k/v store into a document-oriented NoSQL database. Clearly secondary indexes are not yet a very expressive query language, however, they are a major first step towards it.

What does it mean?

The interesting aspect is that Oracle as a major database vendor moved with a separate and dedicated NoSQL product first into the k/v store space, and now additionally into the document-oriented NoSQL space. There are many ways to read this move, of course, but from a technical viewpoint this emphasizes the importance of a structured data model in the NoSQL space. In addition, OracleNoSQL allows a per-document Avro schema, making it a truly per-document schema database.

Aerospike, for example, also supports complex data types and secondary indexes, even though it is also widely seen as a k/v store (e.g., http://db-engines.com/en/ranking/key-value+store).


Clearly, the document-oriented segment of NoSQL databases gets more and more crowded and the attention of major players in the database space.

What hasn’t been done is a comparison of the data model and query interface expressiveness of the major players in the document-oriented NoSQL database segment. Once done, and in case that reveals that the expressiveness is the same or almost the same, then the systems are exchangeable from a functionality viewpoint and will distinguish themselves ‘only’ in terms of non-functional properties like latency, throughput, transactions, cost, support, and so on.

SQL: The Kingmaker of NoSQL Databases?

Here is an interesting question: what if all NoSQL and NewSQL databases would implement SQL?

Advanced SQL

“SQL” in this context does not refer to SQL 92 (http://en.wikipedia.org/wiki/SQL-92) or other versions of the purely relational model focused standard, but to a SQL that is extended properly in syntax and in semantics to deal with non-relational data models. Let’s name it “Advanced SQL” for the purpose of this blog.

I furthermore assume that all data in a database would be accessible, not just parts of it. For example, in key/value stores the values’ internal structure would be accessible as long as it contains data accessible in a structured way (e.g., a JSON representation).

Ideally, of course, all NoSQL and NewSQL databases would implement the same new form of Advanced SQL, but for this blog this is not really necessary or a precondition (but it would be nice, though).

If such an Advanced SQL would be available on all NoSQL and NewSQL databases, then

  • the typical SQL operators like select, project, join, and so on are available on all data in a database across all data models like relational, document-oriented, key/value, columnar, graph-oriented.
  • the same data set loaded into each NoSQL or NewSQL database would be accessible through Advanced SQL and the same query would return the same result, no matter which database (best case assumption wrt. SQL execution semantics).

Data Models

In order to be able to load all data sets into all types of databases the data models have to be equivalent in their structural expressiveness. This might not necessarily be the case today, however, the common trend appears to be that databases are either supporting JSON, or structured data containing array types and nested sub-structures alongside elementary types like string and integer.

For the sake of the blog, let’s assume the data models are equivalent in structural expressiveness.

Implication of Advanced SQL on NoSQL / NewSQL Databases

The biggest implication of Advanced SQL on NoSQL / NewSQL databases would be that all NoSQL / NewSQL databases would be equivalent from a functional perspective in terms of accessing data: a given data set could be stored in any of the databases and queried using SQL. A query would (ideally) return the same result in any of the databases storing the same data set.

The big question this bears is: what would distinguish the NoSQL/NewSQL databases in this case?

  • The non-functional aspects.

For example, to name a few:

  • Performance: the same query would take different execution time
  • Transactions: a set of queries would form a single transaction (or not)
  • Scale: the speed and extent a database can scale up/scale down with varying data set size
  • Backup/restore: how fast and how reliable can databases be re-created after disaster

This basically would mean that a user can select a database based on its non-functional characteristics. (Of course, history has shown that the particular SQL implementation of a database will have proprietary extensions also).

Benefits of Advanced SQL

There are several benefits if Advanced SQL like described above would be available.

  • A user can select not only one, but several databases for different needs, while not having to deal with vastly different query interfaces. For example, one for production system access speed and a different one for long-term storage or analytics.
  • Data can be distributed across different systems depending of how the data is used without having to deal with incompatible access interfaces.
  • NoSQL and NewSQL databases can be integrated alongside existing RDBMSs for mutual benefit.

In a nutshell, users are not forced anymore to select one databases for all their needs, but can select the appropriate one for the particular part of their system architecture.

Trending of Non-functional Features

There is an interesting trend ongoing. OldSQL and NewSQL databases start supporting JSON (e.g., Postgres, VoltDB); Informix has a compatible driver interface to the MongoDB wire format; Tokutek and GridGain implement non-functional replacements for MongoDB; Oracle NoSQL supports per-document Avro schemas for the values; Tokutek adds transaction support; and so on. This looks like the early stages of ‘sorting out’ on a data model level as well as non-functional level. Time will tell if this was the starting point of consolidation in the NoSQL/NewSQL space.

King Making

So why, if the differentiation is likely to happen on the non-functional level, is SQL a potential kingmaker? In my opinion, Advanced SQL will be the standard expectation and a database not supporting Advanced SQL will be relegated to being a niche system. So while the presence of Advanced SQL will be a distinguishing factor (possibly a small one), the absence of Advanced SQL will impact the success of a database significantly.

Relational Data in a Document-oriented NoSQL Database: Overview

This blog starts a series of discussions around the topic of storing relational data in a document-oriented NoSQL database. Each discussion is going to be a separate blog post.

The idea is not to promote storing relational data in a document-oriented database as such. The goal of this series of blogs is to rationalize (to large extent on a blog-level granularity) the relationship between relational and document data and how a relational world can meet a document world (and vice versa).

The starting point of the discussion are the topics in the blog https://realprogrammer.wordpress.com/2012/04/25/relational-database-management-system-rdbms/:

  1. Universal Relation: https://realprogrammer.wordpress.com/2012/05/16/relational-data-in-a-document-oriented-nosql-database-part-1-universal-relation/
  2. Schema: https://realprogrammer.wordpress.com/2012/05/23/relational-data-in-a-document-oriented-nosql-database-part-2-schema/
  3. Normalization: https://realprogrammer.wordpress.com/2012/05/30/relational-data-in-a-document-oriented-nosql-database-part-3-normalization/
  4. Part-Of Relationship: https://realprogrammer.wordpress.com/2012/06/06/relational-data-in-a-document-oriented-nosql-database-part-4-part-of-relationship/
  5. De-Normalization: https://realprogrammer.wordpress.com/2012/06/13/relational-data-in-a-document-oriented-nosql-database-part-5-de-normalization/

More topics might be added as the discussion unfolds.

Based on these discussions I expect that some guidelines will emerge of how to ‘model’ documents in document-oriented databases ‘properly’ and useful criteria for this modeling task. It is also going to be interesting to see if there is a way to determine the relationship between a relational model and a document model in a consistent way.

Modeling documents in a schema-less/dynamic schema world sounds like an oxymoron; however, in the end, transactional applications and analysis software have to access documents and a ‘good’ modeling practice will certainly help those systems in their design and operation.

Computing Technology Inflection Point: Happening Right Now

I believe that computing in industry is at a computing technology inflection point right now based on some supporting observations:

  • Databases. Through the NoSQL movement in the database world a shift is taking place:
    1. New database management systems are being created, made available and put into production
    2. The new database systems are by enlarge schema-less and/or support dynamic schema changes
    3. Many support JSON as data structures or data model that is inherently supported by Javascript
  • Server. The Javascript language is enabled as server-side programming language:
    1. Google’s V8-Engine together with Node.js and its Eco-system makes Javascript a real contender as a server-side language
    2. Javascript does not have a class/instance model and supports schema-less programming, including dynamic schema changes
    3. Integrated Development Environments (IDEs) provide full Javascript support
  • User Interface. The combination of HTML5/Javascript is gaining traction:
    1. HTML5 in conjunction with Javascript is a real powerful web development technology set
    2. Javascript libraries are being built that can span all types of user interfaces and user devices (including mobile)
    3. It is possible to have a streaming query from the database all the way to the user interface; its frictionless

Combining all these observations, two major characteristics of a new computing technology stack are crystallizing very clearly:

  • One cross-layer programming language (Multi-tier Programming)

    • It is possible to use the same language in the User Interface, Server and Database, i.e., across all architectural layers
  • Inherent schema-less computing and dynamic schema change support
    • All layers shift to schema-less technology and models

I believe that this is a real inflection point happening right now and will cause a major shift in system design, engineering and evolution. It leaves behind the notion that each layer in a typical technology stack must have its specialized language. It also leaves behind the notion that every concept across domains should be best represented in the Class-Instance paradigm. It opens up the ability to dynamically evolve within the systems’ schema-less or dynamic schema change capabilities and it supports the reuse of logic and types across all layers without loosing or changing semantics.

If this combination of technology continues to gain major traction, I would not be surprised to see that some of the first operating system prototypes written in Javascript will become mainstream at some point in time. And, of course, it would be interesting to see how far the processor manufacturers are in their thought processes or research projects with putting dedicated Javascript execution functionality on the processors themselves.

Being around at this computing technology inflection points is exciting and professionally lays before us interesting choices in business decisions as well as career decisions.