InfluxDB::Arel

InfluxDB::Arel is a SQL AST manager for InfluxDB dialect. It simplifies the generation of complex SQL queries.

Installation

Add this line to your application's Gemfile:

gem 'influxdb-arel'

And then execute:

$ bundle

Or install it yourself as:

$ gem install influxdb-arel

Usage

Introduction

At start you should create a builder:

builder = InfluxDB::Arel::Builder.new(:events)
builder.select.to_sql
# => SELECT * FROM events

You can set default table name for the builder. Possible to use both strings and symbols:

InfluxDB::Arel::Builder.new('events') == InfluxDB::Arel::Builder.new(:events)
# => true

If you want to use convenient shortcuts, such as 10.h.ago or 1.w you should require a file with core extensions

require 'influx_db/arel/core_extensions'

1.h
# => #<InfluxDB::Arel::Nodes::Duration:0x00000102143a68 @left=1, @right="h">

1.h.to_sql
# => "1h"

1.h.ago.to_sql
# => "(now() - 1h)"

1.h.since.to_sql
# => "(now() + 1h)"

A builder has methods for SQL construction.

• Specifying which attributes should be used in the query: select.

• Specifying which tables and how they should be used in the query: from, merge and join.

• Conditions of query: where.

• Grouping methods: group and fill.

• Ordering methods: order, asc, desc and invert_order.

• Specifying limitations of result set: limit.

• The part of continuous queries: into.

Most of them accept a block for building part of SQL. Inside a block calling of method will be interpreted depending on current context. For example:

In SELECT, WHEREand GROUP contexts:

• All undefined methods will be interpreted as attributes:

builder.where{ pp name.is_a?(InfluxDB::Arel::Nodes::Attribute) }
# true
# => ...
builder.where{ name =~ /undr/ }.to_sql
# => SELECT * FROM table WHERE name =~ /undr/

• Method a returns attribute node.

builder.where{ pp a(:name) == name }
# true
# => ...

• Method time returns InfluxDB::Arel::Nodes::Time object. (It will be available only in GROUP context)

• Method now returns InfluxDB::Arel::Nodes::Now object. (It will be available only in WHERE context)

In FROM, JOIN and MERGE contexts

• All undefined methods will be interpreted as tables:

builder.select{ pp events.is_a?(InfluxDB::Arel::Nodes::Table) }
# true
# => ...
builder.from{ events }.to_sql
# => SELECT * FROM events

• Method t returns table node.

builder.from{ pp t(:table) == table }
# true
# => ...

• Method join used for joining tables (available only in JOIN context).

builder.from{ join(table.as(:alias1), table.as(:alias2)) }.to_sql
# => SELECT * FROM table AS alias1 INNER JOIN table AS alias2

• Method merge used for merging tables (available only in JOIN context).

builder.from{ merge(table1, table2) }.to_sql
# => SELECT * FROM table1 MERGE table2

Also, into the block will be available o method. It used for access to outer scope. For example:

regexp = /events\..*/
builder.from{ o{ regexp } }.to_sql
# => SELECT * FROM /events\..*/

SELECT clause

You can specify attributes or expressions for SELECT clause using select method.

builder = InfluxDB::Arel::Builder.new(:cpu_load)
builder.to_sql
# => SELECT * FROM cpu_load

builder.select{ (system + user).as(:sum) }.to_sql
# => SELECT (system + user) AS sum FROM cpu_load

builder.select{
  [mean(idle).as(:idle_mean), mean(user).as(:user_mean)]
}.group{ time(1.d) }.to_sql
# => SELECT MEAN(idle) AS idle_mean, MEAN(user) AS user_mean FROM cpu_load GROUP BY time(1d)

It might be convenient to use aliases for complex expressions, such as functions or some mathematical expressions.

Chaining select method will add attributes or expressions to the set. If you want to override expressions when use select! method.

builder.select(:name).select(:age).to_sql
# => SELECT name, age FROM table
builder.select(:name).select!(:age).to_sql
# => SELECT age FROM table

FROM clause

You can specify tables for query using from method.

Possible to call method:

• With strings or symbols as arguments

builder.from('events', :errors).to_sql
# => SELECT * FROM events, errors

• With block

builder.from{ errors }.to_sql
# => SELECT * FROM errors

builder.from{ [errors, :events] }.to_sql
# => SELECT * FROM errors, events

• You can mix both

builder.from(:events){ errors }.to_sql
# => SELECT * FROM events, errors

• With regexp object

builder.from(/.*/).to_sql
# => SELECT * FROM /.*/

Warning: You can call method with more then one regexp but only first will be used as table name

builder.from(/.*/, /logs\..*/).to_sql
# => SELECT * FROM /.*/

Chaining this method will replace previous FORM definition.

builder.from(:table1).from{ table2 }.to_sql
# => SELECT * FROM table2

Joining tables

You can join two tables using join method.

It will join two first tables from tables list if method is called without argument

builder = InfluxDB::Arel::Builder.new(:table)
builder.from(:table1, :table2).join.to_sql
# => SELECT * FROM table1 INNER JOIN table2
builder.from{ [table1.as(:alias1), table2.as(:alias2)] }.join.to_sql
# => SELECT * FROM table1 AS alias1 INNER JOIN table2 AS alias2

It will raise error if method is called without argument and tables list contains less than two table.

builder.join.to_sql
# => IllegalSQLConstruct: Ambiguous joining clause

It will join first table from tables list with given table if argument exists.

builder.join(:table2).to_sql
# => SELECT * FROM table INNER JOIN table2

And it will raise error if number of tables is more than two.

builder.from(:table1, :table2).join(:table3).to_sql
# => IllegalSQLConstruct: Ambiguous joining clause

Also, you can define alias for each joined table. It would be useful for self joining table.

builder.from{ table.as(:table_one).join(table.as(:table_two)) }.to_sql
# => SELECT * FROM table AS table_one INNER JOIN table AS table_two

Chaining this method will replace previous join definition.

builder.join(:table1).join{ table1.as(:alias) }.to_sql
# => SELECT * FROM table INNER JOIN table1 AS alias

Merging tables

You can merge two tables using merge method.

It will merge two first tables from tables list if method is called without argument

builder = InfluxDB::Arel::Builder.new(:table)
builder.from(:table1, :table2).merge.to_sql
# => SELECT * FROM table1 MERGE table2
builder.from{ [table1.as(:alias1), table2.as(:alias2)] }.merge.to_sql
# => SELECT * FROM table1 MERGE table2

It will raise error if method is called without argument and tables list contains less than two table.

builder.join.to_sql
# => IllegalSQLConstruct: Ambiguous merging clause

It will merge first table from tables list with given table if argument exists.

builder.merge(:table2).to_sql
# => SELECT * FROM table MERGE table2

And it will raise error if number of tables is more than two.

builder.from(:table1, :table2).merge(:table3).to_sql
# => IllegalSQLConstruct: Ambiguous merging clause

Also, you can define alias for each table, but it doesn't influence on result. Because aliases make sense only for joining tables.

builder.from{ table1.as(:table1).merge(table1.as(:table2)) }.to_sql
# => SELECT * FROM table1 MERGE table2

Chaining this method will replace previous merge definition.

builder.merge(:table1).merge(:table2).to_sql
# => SELECT * FROM table MERGE table2

Grouping of results

Grouping of results by specified attributes or expressions, such as time(10m):

builder = InfluxDB::Arel::Builder.new(:table)
builder.group{ [time(10.m), host] }.to_sql
# => SELECT * FROM table GROUP BY time(10m), host

If you want to fill intervals with no data you should call fill method:

builder.group{ [time(10.m), host] }.fill(0).to_sql
# => SELECT * FROM table GROUP BY time(10m), host fill(0)

Chaining this method will add expression to the grouping set. If you want to override expressions when use group! method.

builder.group{ time(10.m) }.group(:host).to_sql
# => SELECT * FROM table GROUP BY time(10m), host
builder.group{ time(10.m) }.group!(:host).to_sql
# => SELECT * FROM table GROUP BY host

Ordering of results

Yo can set the ordering of results using order method

Possible values:

• :asc- Default value. Results will be sorted by ascending order.
• 'asc'- Results will be sorted by ascending order.
• :desc- Results will be sorted by descending order.
• 'desc'- Results will be sorted by descending order.

builder = InfluxDB::Arel::Builder.new(:table)
builder.order(:desc).to_sql
builder.order('desc').to_sql
# => SELECT * FROM table ORDER DESC

builder.order(:asc).to_sql
builder.order('asc').to_sql
# => SELECT * FROM table ORDER ASC

As well it's possible to use asc and desc methods

builder.asc.to_sql
# => SELECT * FROM table ORDER ASC
builder.desc.to_sql
# => SELECT * FROM table ORDER DESC

Inverting of the order direction:

builder.asc.invert_order.to_sql
# => SELECT * FROM table ORDER DESC
builder.desc.invert_order.to_sql
# => SELECT * FROM table ORDER ASC

Chaining this methods will replace previous order definition.

builder.asc.desc.to_sql
# => SELECT * FROM table ORDER DESC

Query conditions

You can specify conditions for query using where method

builder = InfluxDB::Arel::Builder.new(:table)
builder.where(name: 'Undr').to_sql
# => SELECT * FROM table WHERE name = 'Undr'

builder.where{ name.eq('Undr').or(name.eq('Andrei')) }.to_sql
# => SELECT * FROM table WHERE name = 'Undr' OR name = 'Andrei'

Chaining this method will add expression to the condition set. They will join using AND boolean expression. If you want to override conditions when use where! method.

builder.where(name: 'Undr').where{ time.lt(10.h.ago) }.to_sql
# => SELECT * FROM table WHERE name = 'Undr' AND time < (now() - 10h)
builder.where(name: 'Undr').where!{ time.lt(10.h.ago) }.to_sql
# => SELECT * FROM table WHERE time < (now() - 10h)

Limits

You can set a limit for a result set

builder = InfluxDB::Arel::Builder.new(:cpu_load)
builder.limit(100).to_sql
# => SELECT * FROM cpu_load LIMIT 100

Contributing

1. Fork it ( https://github.com/undr/influxdb-arel/fork )
2. Create your feature branch (git checkout -b my-new-feature)
3. Commit your changes (git commit -am 'Add some feature')
4. Push to the branch (git push origin my-new-feature)
5. Create a new Pull Request