Unify disparate data sources and files into a single Federated view of your data and query with SQL without copying into datawarehouse.

Mysql compatible federated query engine to Elasticsearch, Mongo, Google Datastore, Cassandra, Google BigTable, Kubernetes, file-based sources. This query engine hosts a mysql protocol listener, which rewrites sql queries to native (elasticsearch, mongo, cassandra, kuberntes-rest-api, bigtable). It works by implementing a full relational algebra distributed execution engine to run sql queries and poly-fill missing features from underlying sources. So, a backend key-value storage such as cassandra can now have complete WHERE clause support as well as aggregate functions etc.

Most similar to prestodb but in Golang, and focused on easy to add custom data sources as well as REST api sources.

• Google Big Table SQL against big-table Bigtable.
• Elasticsearch Simplify access to Elasticsearch.
• Mongo Translate SQL into mongo.
• Google Cloud Storage / (csv, json files) An example of REST api backends (list of files), as well as the file contents themselves are tables.
• Cassandra SQL against cassandra. Adds sql features that are missing.
• Lytics SQL against Lytics REST Api's
• Kubernetes An example of REST api backend.
• Google Big Query MYSQL against worlds best analytics datawarehouse BigQuery.
• Google Datastore MYSQL against Datastore.

• Distributed run queries across multiple servers
• Hackable Sources Very easy to add a new Source for your custom data, files, json, csv, storage.
• Hackable Functions Add custom go functions to extend the sql language.
• Joins Get join functionality between heterogeneous sources.
• Frontends currently only MySql protocol is supported but RethinkDB (for real-time api) is planned, and are pluggable.
• Backends Elasticsearch, Google-Datastore, Mongo, Cassandra, BigTable, Kubernetes currently implemented. Csv, Json files, and custom formats (protobuf) are in progress.

• NOT Production ready. Currently supporting a few non-critical use-cases (ad-hoc queries, support tool) in production.

This example imports a couple hours worth of historical data from https://www.githubarchive.org/ into a local elasticsearch server for example. Requires etcd server running local as well. Docker setup coming soon.

# assuming you are running local, if you are instead in Google Cloud, or Google Container Engine
# you don't need the credentials or volume mount
docker run -e "GOOGLE_APPLICATION_CREDENTIALS=/.config/gcloud/application_default_credentials.json" \
  -e "LOGGING=debug" \
  --rm -it \
  -p 4000:4000 \
  -v ~/.config/gcloud:/.config/gcloud \
  gcr.io/dataux-io/dataux:latest

# now that dataux is running use mysql-client to connect
mysql -h 127.0.0.1 -P 4000

now run some queries

-- we have a default CSV dataset that we have copied to the local
-- docker drive from http://seanlahman.com/baseball-archive/statistics/ 

show databases;

use baseball;

show tables;

describe appearances

select count(*) from appearances;

select * from appearances limit 10;


-- Lets add a bigquery datasource
CREATE source `datauxtest` WITH {
    "type":"bigquery",
    "schema":"bqsf_bikes",
    "table_aliases" : {
       "bikeshare_stations" : "bigquery-public-data:san_francisco.bikeshare_stations"
    },
    "settings" : {
      "billing_project" : "your-google-cloud-project",
      "data_project" : "bigquery-public-data",
      "dataset" : "san_francisco"
    }
};

use bqsf_bikes;

show tables;

describe film_locations;

select * from film_locations limit 10;

• Sources
  • Big-Query
  • Big-Table
  • Json-Files
• Writes
  • write pub/sub: inbound insert/update are available as pub-sub messages.
  • write propogation: inbound insert/update gets written multiple places.
  • write lambda functions: allow arbitray functions to get nats.io pub/sub of write events.

Hacking

• see tools/importgithub for tool to import 2 days of github data for examples above.

# to run queries below, run test data import

go get -u github.com/dataux/dataux
cd $GOPATH/src/github.com/dataux/dataux/tools/importgithub
go build
./importgithub  ## will import ~200k plus docs from Github archive

• Data-Accessability Making it easier to query, access, share, and use data. Protocol shifting (for accessibility). Sharing/Replication between db types.
• Scalability/Sharding Implement sharding, connection sharing

We use more and more databases, flatfiles, message queues, etc. For db's the primary reader/writer is fine but secondary readers such as investigating ad-hoc issues means we might be accessing and learning many different query languages.

Credit to mixer, derived mysql connection pieces from it (which was forked from vitess).

• https://github.com/linkedin/databus,
• ql.io, yql
• dockersql, q -python, textql,GitQL/GitQL, GitQL

In Internet architectures, data systems are typically categorized into source-of-truth systems that serve as primary stores for the user-generated writes, and derived data stores or indexes which serve reads and other complex queries. The data in these secondary stores is often derived from the primary data through custom transformations, sometimes involving complex processing driven by business logic. Similarly data in caching tiers is derived from reads against the primary data store, but needs to get invalidated or refreshed when the primary data gets mutated. A fundamental requirement emerging from these kinds of data architectures is the need to reliably capture, flow and process primary data changes.

from Databus