Batch statements

Session objects can be used to construct a logged batch statement and later execute it.

Background

Given

a running cassandra cluster with schema:

CREATE KEYSPACE simplex WITH replication = {'class': 'SimpleStrategy', 'replication_factor': 3};
CREATE TABLE simplex.songs (
  id uuid PRIMARY KEY,
  title text,
  album text,
  artist text,
  tags set<text>
);
CREATE TABLE simplex.cas_batch (k text, v int, PRIMARY KEY (k, v));

since cassadra v2.0

A batch of simple statements is executed

Given

the following example:

require 'cassandra'

cluster = Cassandra.cluster
session = cluster.connect("simplex")

rows = session.execute("SELECT * FROM songs")

puts "songs contain #{rows.size} rows"

batch   = session.batch do |b|
            b.add("INSERT INTO songs (id, title, album, artist, tags)
                   VALUES (
                      756716f7-2e54-4715-9f00-91dcbea6cf50,
                      'La Petite Tonkinoise',
                      'Bye Bye Blackbird',
                      'Joséphine Baker',
                      {'jazz', '2013'}
                   )")
            b.add("INSERT INTO songs (id, title, album, artist, tags)
                   VALUES (
                      f6071e72-48ec-4fcb-bf3e-379c8a696488,
                      'Die Mösch',
                      'In Gold',
                      'Willi Ostermann',
                      {'kölsch', '1996', 'birds'}
                   )")
            b.add("INSERT INTO songs (id, title, album, artist, tags)
                   VALUES (
                      fbdf82ed-0063-4796-9c7c-a3d4f47b4b25,
                      'Memo From Turner',
                      'Performance',
                      'Mick Jager',
                      {'soundtrack', '1991'}
                   )")
          end

puts "inserting rows in a batch"

session.execute(batch, consistency: :all)
rows = session.execute("SELECT * FROM songs")

puts "songs contain #{rows.size} rows"

When

it is executed

Then

its output should contain:

songs contain 0 rows
inserting rows in a batch
songs contain 3 rows

since cassadra v2.0

A batch of simple statements with parameters is executed

Given

the following example:

require 'cassandra'

cluster = Cassandra.cluster
session = cluster.connect("simplex")

rows = session.execute("SELECT * FROM songs")

puts "songs contain #{rows.size} rows"

batch   = session.batch do |b|
            b.add("INSERT INTO songs (id, title, album, artist, tags) VALUES (?, ?, ?, ?, ?)",
                  arguments: [ Cassandra::Uuid.new('756716f7-2e54-4715-9f00-91dcbea6cf50'),
                               'La Petite Tonkinoise',
                               'Bye Bye Blackbird',
                               'Joséphine Baker',
                               Set['jazz', '2013']
                             ]
            )
            b.add("INSERT INTO songs (id, title, album, artist, tags) VALUES (?, ?, ?, ?, ?)",
                  arguments: [ Cassandra::Uuid.new('f6071e72-48ec-4fcb-bf3e-379c8a696488'),
                               'Die Mösch',
                               'In Gold',
                               'Willi Ostermann',
                               Set['kölsch', '1996', 'birds']
                             ]
            )
            b.add("INSERT INTO songs (id, title, album, artist, tags) VALUES (?, ?, ?, ?, ?)",
                  arguments: [ Cassandra::Uuid.new('fbdf82ed-0063-4796-9c7c-a3d4f47b4b25'),
                               'Memo From Turner',
                               'Performance',
                               'Mick Jager',
                               Set['soundtrack', '1991']
                             ]
            )
          end

puts "inserting rows in a batch"

session.execute(batch, consistency: :all)
rows = session.execute("SELECT * FROM songs")

puts "songs contain #{rows.size} rows"

When

it is executed

Then

its output should contain:

songs contain 0 rows
inserting rows in a batch
songs contain 3 rows

since cassadra v2.0

A prepared statement is executed in a batch

Given

the following example:

require 'cassandra'

cluster = Cassandra.cluster
session = cluster.connect("simplex")

rows = session.execute("SELECT * FROM songs")

puts "songs contain #{rows.size} rows"

insert  = session.prepare("INSERT INTO songs (id, title, album, artist, tags) VALUES (?, ?, ?, ?, ?)")
batch   = session.batch do |b|
            b.add(insert,
                  arguments: [ Cassandra::Uuid.new('756716f7-2e54-4715-9f00-91dcbea6cf50'),
                               'La Petite Tonkinoise',
                               'Bye Bye Blackbird',
                               'Joséphine Baker',
                               Set['jazz', '2013']
                             ]
            )
            b.add(insert,
                  arguments: [ Cassandra::Uuid.new('f6071e72-48ec-4fcb-bf3e-379c8a696488'),
                               'Die Mösch',
                               'In Gold',
                               'Willi Ostermann',
                               Set['kölsch', '1996', 'birds']
                             ]
            )
            b.add(insert,
                  arguments: [ Cassandra::Uuid.new('fbdf82ed-0063-4796-9c7c-a3d4f47b4b25'),
                               'Memo From Turner',
                               'Performance',
                               'Mick Jager',
                               Set['soundtrack', '1991']
                             ]
            )
          end

puts "inserting rows in a batch"

session.execute(batch, consistency: :all)
rows = session.execute("SELECT * FROM songs")

puts "songs contain #{rows.size} rows"

When

it is executed

Then

its output should contain:

songs contain 0 rows
inserting rows in a batch
songs contain 3 rows

since cassadra v2.1

A prepared statement is executed in a batch with named parameters

Given

the following example:

  require 'cassandra'

  cluster = Cassandra.cluster
  session = cluster.connect("simplex")

  rows = session.execute("SELECT * FROM songs")

  puts "songs contain #{rows.size} rows"

  insert  = session.prepare("INSERT INTO songs (id, title, album, artist, tags) VALUES (:a, :b, :c, :d, :e)")
  batch   = session.batch do |b|
              b.add(insert, arguments: {
                    :a => Cassandra::Uuid.new('756716f7-2e54-4715-9f00-91dcbea6cf50'),
                    :b => 'La Petite Tonkinoise',
                    :c =>'Bye Bye Blackbird',
                    :d => 'Joséphine Baker',
                    :e => Set['jazz', '2013']
              })
              b.add(insert, arguments: {
                    :a => Cassandra::Uuid.new('f6071e72-48ec-4fcb-bf3e-379c8a696488'),
                    :b => 'Die Mösch',
                    :c => 'In Gold',
                    :d => 'Willi Ostermann',
                    :e => Set['kölsch', '1996', 'birds']
              })
              b.add(insert, arguments: {
                    :a => Cassandra::Uuid.new('fbdf82ed-0063-4796-9c7c-a3d4f47b4b25'),
                    :b => 'Memo From Turner',
                    :c => 'Performance',
                    :d => 'Mick Jager',
                    :e => Set['soundtrack', '1991']
              })
            end

  puts "inserting rows in a batch"

  session.execute(batch, consistency: :all)
  rows = session.execute("SELECT * FROM songs")

  puts "songs contain #{rows.size} rows"

When

it is executed

Then

its output should contain:

songs contain 0 rows
inserting rows in a batch
songs contain 3 rows

since cassadra v2.0.9

A cas batch is never applied more than once

Given

the following example:

require 'cassandra'

cluster = Cassandra.cluster
session = cluster.connect("simplex")

statement = session.prepare("INSERT INTO cas_batch (k, v) VALUES (?, ?) IF NOT EXISTS")
batch = session.batch

batch.add("INSERT INTO cas_batch (k, v) VALUES ('key1', 0)")
batch.add(statement, arguments: ["key1", 1])
batch.add(statement, arguments: ["key1", 2])

results =  session.execute(batch)
rows = results.first
puts "batch applied? #{rows["[applied]"]}"

results =  session.execute(batch)
rows = results.first
puts "batch applied? #{rows["[applied]"]}"

When

it is executed

Then

its output should contain:

batch applied? true
batch applied? false

since cassadra v1.2

Cassandra 1.2 doesn’t support batch statements

Given

the following example:

require 'cassandra'

cluster = Cassandra.cluster
session = cluster.connect("simplex")
batch   = session.batch

batch.add("INSERT INTO cas_batch (k, v) VALUES ('key1', 0)")

begin
  session.execute(batch)
rescue => e
  puts "#{e.class.name}: #{e.message}"
end

When

it is executed

Then

its output should contain:

Cassandra::Errors::ClientError: Batch statements are not supported by the current version of Apache Cassandra

since cassadra v2.1

A batch write is executed with serial_consistency set

Given

the following example:

  require 'cassandra'

  cluster = Cassandra.cluster
  session = cluster.connect("simplex")

  rows = session.execute("SELECT * FROM songs")
  puts "songs contain #{rows.size} rows"

  batch   = session.batch do |b|
              b.add("INSERT INTO songs (id, title, album, artist, tags)
                     VALUES (
                        756716f7-2e54-4715-9f00-91dcbea6cf50,
                        'La Petite Tonkinoise',
                        'Bye Bye Blackbird',
                        'Joséphine Baker',
                        {'jazz', '2013'}
                     )")
            end

  puts "inserting rows in a batch"
  session.execute(batch, consistency: :all)
  row = session.execute("SELECT * FROM songs").first
  puts "song name: #{row['title']}"

  batch = session.batch do |b|
            b.add("UPDATE songs SET
                      title  = 'Die Mösch',
                      album  = 'In Gold',
                      artist = 'Willi Ostermann',
                      tags   = {'kölsch', '1996', 'birds'}
                    WHERE id = 756716f7-2e54-4715-9f00-91dcbea6cf50
                    IF title = 'La Petite Tonkinoise'
                  ")
           end

  puts "updating rows in a batch"
  session.execute(batch, consistency: :all, serial_consistency: :serial)
  row = session.execute("SELECT * FROM songs").first
  puts "song name: #{row['title']}"

  update = session.prepare("UPDATE songs SET
                              title  = 'Memo From Turner',
                              album  = 'Performance',
                              artist = 'Mick Jager',
                              tags   = {'soundtrack', '1991'}
                            WHERE id = ?
                            IF title = 'Die Mösch'
                            ")

  batch = session.batch do |b|
            b.add(update, arguments: [Cassandra::Uuid.new('756716f7-2e54-4715-9f00-91dcbea6cf50')])
          end

  puts "updating rows in a batch"
  session.execute(batch, consistency: :all, serial_consistency: :local_serial)
  row = session.execute("SELECT * FROM songs").first
  puts "song name: #{row['title']}"

When

it is executed

Then

its output should contain:

songs contain 0 rows
inserting rows in a batch
song name: La Petite Tonkinoise
updating rows in a batch
song name: Die Mösch
updating rows in a batch
song name: Memo From Turner

Batch statements

Background

A batch of simple statements is executed

A batch of simple statements with parameters is executed

A prepared statement is executed in a batch

A prepared statement is executed in a batch with named parameters

A cas batch is never applied more than once

Cassandra 1.2 doesn’t support batch statements

A batch write is executed with serial_consistency set

Contents