Making Data Dance (PGConfSV)

Your web framework and ORM are really handy tools to get work done, and do it quickly. They aren't a silver bullet, though. Sometimes the database is the best place to run a complicated query. Especially when you have big transactional or time-series data. Postgres has some powerful functionality that can help transform data from a giant pile of stuff into something actionable and useful. I will show how to take advantage of the JSON datatype, Common Table Expression, and materialized views to build impressive data visualizations.

Transcript

1. MAKING DATA DANCE PGConf Silicon Valley @barrettclark
2. HARNESS HERETICAL ANTIPATTERNS FOR FUN AND PROFIT PGConf Silicon Valley @barrettclark
3. BARRETT CLARK @barrettclark PGConf Silicon Valley @barrettclark
4. PGConf Silicon Valley @barrettclark
5. Class (Model) Object Relational Mapper Database M agic PGConf Silicon Valley @barrettclark
6. DATABASE AGNOSTIC PGConf Silicon Valley @barrettclark
7. ActiveRecord Object Relational Mapper PGConf Silicon Valley @barrettclark
8. ActiveRecord PGConf Silicon Valley @barrettclark
9. RAW SQL PGConf Silicon Valley @barrettclark
10. COMPLICATED JOINS PGConf Silicon Valley @barrettclark
11. Category.joins(articles: [{ comments: :guest }, :tags]) http://guides.rubyonrails.org/active_record_querying.html#using-array-hash-of-named-associations ??? PGConf Silicon Valley @barrettclark
12. SELECT categories.* FROM categories INNER JOIN articles ON articles.category_id = categories.id INNER JOIN comments ON comments.article_id = articles.id INNER JOIN guests ON guests.comment_id = comments.id INNER JOIN tags ON tags.article_id = articles.id http://guides.rubyonrails.org/active_record_querying.html#using-array-hash-of-named-associations Category.joins(articles: [{ comments: :guest }, :tags]) PGConf Silicon Valley @barrettclark
13. DEEPER DATABASE FUNCTIONALITY PGConf Silicon Valley @barrettclark
14. BUT DATABASE PORTABILITY?! PGConf Silicon Valley @barrettclark
15. https://flic.kr/p/eg4hih PGConf Silicon Valley @barrettclark
16. It just doesn't work like that. Database portability pipe dream PGConf Silicon Valley @barrettclark
17. Different databases are different. PGConf Silicon Valley @barrettclark
18. You chose your database for a reason. Use it! PGConf Silicon Valley @barrettclark
19. CAVEATS • Long-running queries • Connection pool PGConf Silicon Valley @barrettclark
20. LET'S GET OUR HANDS DIRTY PGConf Silicon Valley @barrettclark
21. PGConf Silicon Valley @barrettclark
22. WINDOW FUNCTIONS PGConf Silicon Valley @barrettclark
23. –Postgres Documentation A window function performs a calculation across a set of table rows that are somehow related to the current row. PGConf Silicon Valley @barrettclark
24. –Barrett Clark Fold data from other rows into this row PGConf Silicon Valley @barrettclark
25. WINDOW FUNCTION GREATEST HITS • lead() • lag() • first_value() • last_value() • row_number() PGConf Silicon Valley @barrettclark
26. id fruit 1 apple 2 apple 3 apple 4 pear 5 apple 6 pear 7 pear 8 pear 9 pear 10 banana PGConf Silicon Valley @barrettclark
27. SELECT *,LEAD(id) OVER(), LAG(id) OVER() FROM fruits; id | fruit | lead | lag ----+--------+------+----- 1 | apple | 2 | 2 | apple | 3 | 1 3 | apple | 4 | 2 4 | pear | 5 | 3 5 | apple | 6 | 4 6 | pear | 7 | 5 7 | pear | 8 | 6 8 | pear | 9 | 7 9 | pear | 10 | 8 10 | banana | | 9 (10 rows) http://sqlfiddle.com/#!15/230c3b/2 PGConf Silicon Valley @barrettclark
28. –Postgres Documentation A window function performs a calculation across a set of table rows that are somehow related to the current row. PGConf Silicon Valley @barrettclark
29. https://flic.kr/p/7KGqEY PGConf Silicon Valley @barrettclark
30. SELECT id, fruit, FIRST_VALUE(id) OVER(PARTITION BY fruit ORDER BY id) FROM fruits; id | fruit | first_value ----+--------+------------- 1 | apple | 1 2 | apple | 1 3 | apple | 1 5 | apple | 1 10 | banana | 10 4 | pear | 4 6 | pear | 4 7 | pear | 4 8 | pear | 4 9 | pear | 4 (10 rows) http://sqlfiddle.com/#!15/230c3b PGConf Silicon Valley @barrettclark
31. SELECT *, FIRST_VALUE(id) OVER(PARTITION BY fruit ORDER BY id), ROW_NUMBER() OVER(PARTITION BY fruit ORDER BY id) FROM fruits; id | fruit | first_value | row_number ----+--------+-------------+------------ 1 | apple | 1 | 1 2 | apple | 1 | 2 3 | apple | 1 | 3 5 | apple | 1 | 4 10 | banana | 10 | 1 4 | pear | 4 | 1 6 | pear | 4 | 2 7 | pear | 4 | 3 8 | pear | 4 | 4 9 | pear | 4 | 5 (10 rows) http://sqlfiddle.com/#!15/230c3b PGConf Silicon Valley @barrettclark
32. PRACTICAL EXAMPLE Chat App PGConf Silicon Valley @barrettclark
33. SELECT id, message, room_id, user_id FROM messages ORDER BY id; id | message | room_id | user_id ----+-----------------------------+---------+--------- 1 | Leaving for the airport now | 1 | 1337 2 | Flight just landed! | 1 | 1337 3 | waiting on an uber | 2 | 1337 4 | omw y'all! | 1 | 1337 9 | hello, world | 1 | 27 (5 rows) PGConf Silicon Valley @barrettclark
34. SELECT id, message, room_id, user_id, LAG(user_id) OVER( PARTITION BY room_id ORDER BY room_id, id ) AS prev_user_id FROM messages ORDER BY 3, 1; Look back Defines the groups of rows to evaluate against the current row PGConf Silicon Valley @barrettclark
35. id | message | room_id | user_id | prev_user_id ----+-----------------------------+---------+---------+-------------- 1 | Leaving for the airport now | 1 | 1337 | 2 | Flight just landed! | 1 | 1337 | 1337 4 | omw y'all! | 1 | 1337 | 1337 9 | hello, world | 1 | 27 | 1337 3 | waiting on an uber | 2 | 1337 | (5 rows) SELECT id, message, room_id, user_id, LAG(user_id) OVER( PARTITION BY room_id ORDER BY room_id, id ) AS prev_user_id FROM messages ORDER BY 3, 1; PGConf Silicon Valley @barrettclark
36. SELECT id, message, room_id, user_id, LAG(user_id) OVER( PARTITION BY room_id ORDER BY room_id, id ) AS prev_user_id FROM messages ORDER BY 3, 1; ProTip™ 3 1 2 4 5 PGConf Silicon Valley @barrettclark
37. class Message < ActiveRecord::Base def self.window_example sql = <<-SQL SELECT id, message, room_id, user_id, LAG(user_id) OVER(PARTITION BY room_id ORDER BY room_id, id) AS prev_user_id FROM messages ORDER BY 3, 1; SQL connection.execute(sql) end end PGConf Silicon Valley @barrettclark
38. TOO ANTI-PATTERN FOR YOU? PGConf Silicon Valley @barrettclark
39. module PGConnection def conn config = YAML.load_file(File.open('config/database.yml'))['development'] @conn ||= PG.connect( :dbname => config['database'], :user => config['username'], :password => config['password'], :host => config['host'] || 'localhost' ) end end class Antipattern extend PGConnection def self.window_example sql = <<-SQL SELECT id, message, room_id, user_id, LAG(user_id) OVER(PARTITION BY room_id ORDER BY room_id, id) AS prev_user_id FROM messages ORDER BY 3, 1; SQL conn.exec(sql).values end end PGConf Silicon Valley @barrettclark
40. MODELS DO NOT HAVE TO BE BACKED BY ACTIVERECORD Or even persisted in the database. PGConf Silicon Valley @barrettclark
41. QUESTIONS WE CAN NOW ANSWER • When did something change? • When did someone leave a place? • How long did each thing last? • How long did people stay in which places? • Enter / Exit / Change events PGConf Silicon Valley @barrettclark
42. https://flic.kr/p/5b8cq3 PGConf Silicon Valley @barrettclark
43. SUBQUERIES PGConf Silicon Valley @barrettclark
44. SUBQUERY • To filter and group • For a field value • In a join PGConf Silicon Valley @barrettclark
45. BASE QUERY PGConf Silicon Valley @barrettclark
46. SELECT id, phone, major, minor, reading_timestamp, LEAD(minor) OVER ( PARTITION BY phone ORDER BY phone, reading_timestamp ) AS next_minor FROM beacon_readings ORDER BY 2, 5; “For a given phone, pull in the next minor field” PGConf Silicon Valley @barrettclark
47. WHEN DID SOMETHING CHANGE? PGConf Silicon Valley @barrettclark
48. SELECT major, minor, next_minor, COUNT(*) FROM ( SELECT id, phone, major, minor, reading_timestamp, LEAD(minor) OVER ( PARTITION BY phone ORDER BY phone, reading_timestamp ) AS next_minor FROM beacon_readings ORDER BY 2, 5 ) AS beacon_readings_lead WHERE major = 1 AND minor != next_minor GROUP BY major, minor, next_minor ORDER BY major, minor, next_minor; PGConf Silicon Valley @barrettclark
49. YO DAWG, I HEARD YOU LIKE QUERIES PGConf Silicon Valley @barrettclark
50. SELECT major, minor, next_minor, COUNT(*) FROM ( SELECT id, phone, major, minor, reading_timestamp, LEAD(minor) OVER ( PARTITION BY phone ORDER BY phone, reading_timestamp ) AS next_minor FROM beacon_readings ORDER BY 2, 5 ) AS beacon_readings_lead WHERE major = 1 AND minor != next_minor GROUP BY major, minor, next_minor ORDER BY major, minor, next_minor; PGConf Silicon Valley @barrettclark
51. major | minor | next_minor | count -------+-------+------------+------- 1 | 0 | 1 | 18 1 | 0 | 2 | 3 1 | 0 | 3 | 23 1 | 0 | 4 | 7 1 | 1 | 0 | 19 1 | 1 | 2 | 59 1 | 1 | 3 | 158 1 | 1 | 4 | 44 1 | 2 | 0 | 5 1 | 2 | 1 | 59 1 | 2 | 3 | 85 1 | 2 | 4 | 40 1 | 3 | 0 | 21 1 | 3 | 1 | 154 1 | 3 | 2 | 85 1 | 3 | 4 | 52 1 | 4 | 0 | 6 1 | 4 | 1 | 49 1 | 4 | 2 | 42 1 | 4 | 3 | 45 (20 rows) PGConf Silicon Valley @barrettclark
52. PGConf Silicon Valley @barrettclark
53. https://flic.kr/p/9tT9Yh PGConf Silicon Valley @barrettclark
54. REFACTOR PGConf Silicon Valley @barrettclark
55. CTE Common Table Expression PGConf Silicon Valley @barrettclark
56. SELECT major, minor, next_minor, COUNT(*) FROM ( SELECT id, phone, major, minor, reading_timestamp, LEAD(minor) OVER ( PARTITION BY phone ORDER BY phone, reading_timestamp ) AS next_minor FROM beacon_readings ORDER BY 2, 5 ) AS beacon_readings_lead WHERE major = 1 AND minor != next_minor GROUP BY major, minor, next_minor ORDER BY major, minor, next_minor; PGConf Silicon Valley @barrettclark
57. WITH beacon_readings_lead AS ( SELECT id, phone, major, minor, reading_timestamp, LEAD(minor) OVER ( PARTITION BY phone ORDER BY phone, reading_timestamp ) AS next_minor FROM beacon_readings ORDER BY 2, 5 ) SELECT major, minor, next_minor, COUNT(*) FROM beacon_readings_lead WHERE major = 1 AND minor != next_minor GROUP BY major, minor, next_minor ORDER BY major, minor, next_minor; PGConf Silicon Valley @barrettclark
58. WITH interval_query AS ( SELECT (ts ||' hour')::INTERVAL AS hour_interval FROM generate_series(0,23) AS ts ), time_series AS ( SELECT DATE_TRUNC('hour', NOW()) + INTERVAL '60 min' * ROUND(DATE_PART('minute', NOW()) / 60.0) - interval_query.hour_interval AS start_time FROM interval_query ), time_intervals AS ( SELECT start_time, start_time + '1 hour'::INTERVAL AS end_time FROM time_series ORDER BY start_time ), reading_counts AS ( SELECT f.start_time, f.end_time, br.minor, COUNT(DISTINCT br.phone) readings FROM beacon_readings br RIGHT JOIN time_intervals f ON br.reading_timestamp >= f.start_time AND br.reading_timestamp < f.end_time AND br.major = 1 GROUP BY f.start_time, f.end_time, br.minor ORDER BY f.start_time, br.minor ) SELECT * FROM reading_counts; PGConf Silicon Valley @barrettclark
59. PGConf Silicon Valley @barrettclark
60. Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition Repetition PGConf Silicon Valley @barrettclark
61. VIEW Stored Query PGConf Silicon Valley @barrettclark
62. CREATE OR REPLACE VIEW beacon_readings_lead_view AS SELECT id, phone, major, minor, reading_timestamp, LEAD(minor) OVER ( PARTITION BY phone ORDER BY phone, reading_timestamp ) AS next_minor FROM beacon_readings ORDER BY 2, 5 ; PGConf Silicon Valley @barrettclark
63. class CreateLeadView < ActiveRecord::Migration def up execute <<-SQL.strip_heredoc CREATE OR REPLACE VIEW beacon_readings_lead_view AS SELECT ID, phone, major, minor, reading_timestamp, LEAD(minor) OVER ( PARTITION BY phone ORDER BY phone, reading_timestamp ) AS next_minor FROM beacon_readings ORDER BY phone, reading_timestamp; SQL end def down sql = "DROP VIEW IF EXISTS beacon_readings_lead_view;" execute(sql) end end PGConf Silicon Valley @barrettclark
64. SELECT major, minor, next_minor, COUNT(*) FROM beacon_readings_lead_view WHERE major = 1 AND minor != next_minor GROUP BY major, minor, next_minor ORDER BY major, minor, next_minor; PGConf Silicon Valley @barrettclark
65. BUSINESS LOGIC IN THE DATABASE?!? PGConf Silicon Valley @barrettclark
66. YOU Get to decide where your code runs PGConf Silicon Valley @barrettclark
67. https://flic.kr/p/2MW7SY PGConf Silicon Valley @barrettclark
68. MY QUERY IS TOO SLOW PGConf Silicon Valley @barrettclark
69. DO THE HEAVY LIFTING IN A SEPARATE PROCESS PGConf Silicon Valley @barrettclark
70. MATERIALIZED VIEW Stored query AND data PGConf Silicon Valley @barrettclark
71. CREATE MATERIALIZED VIEW beacon_readings_lead_lag_mv AS SELECT * FROM beacon_readings_lead_lag_view WHERE reading_timestamp >= NOW() - '1 day'::INTERVAL; CREATE INDEX index_beacon_readings_mv_on_major ON beacon_readings_lead_lag_mv USING btree( major ASC NULLS LAST ); CREATE INDEX index_beacon_readings_mv_on_phone ON beacon_readings_lead_lag_mv USING btree( phone COLLATE "default" ASC NULLS LAST ); CREATE INDEX index_beacon_readings_mv_on_reading_timestamp ON beacon_readings_lead_lag_mv USING btree( reading_timestamp ASC NULLS LAST ); PGConf Silicon Valley @barrettclark
72. DATA https://flic.kr/p/khS6jn PGConf Silicon Valley @barrettclark
73. REFRESH MATERIALIZED VIEW CONCURRENTLY beacon_readings_lead_lag_mv; Available starting in 9.4 PGConf Silicon Valley @barrettclark
74. namespace :db do namespace :heroku do # https://devcenter.heroku.com/articles/scheduler desc 'Update the materialized view(s)' task :update_materialized_view => :environment do sql = 'REFRESH MATERIALIZED VIEW beacon_readings_lead_lag_mv;' ActiveRecord::Base.connection.execute(sql) end end end PGConf Silicon Valley @barrettclark
75. https://flic.kr/p/bVUL7 PGConf Silicon Valley @barrettclark
76. SUBQUERY FOR A FIELD VALUE PGConf Silicon Valley @barrettclark
77. SELECT id, phone, ( SELECT MIN(id) FROM beacon_readings_lead_lag_mv WHERE major = br1.major AND minor = 0 AND phone = br1.phone AND id >= br1.id ) AS session_close_id, minor, next_beacon_minor, reading_timestamp FROM beacon_readings_lead_lag_mv br1 WHERE major = 2 AND minor != next_beacon_minor ORDER BY phone, reading_timestamp; PGConf Silicon Valley @barrettclark
78. SELECT id, phone, ( SELECT MIN(id) FROM beacon_readings_lead_lag_mv WHERE major = br1.major AND minor = 0 AND phone = br1.phone AND id >= br1.id ) AS session_close_id, minor, next_beacon_minor, reading_timestamp FROM beacon_readings_lead_lag_mv br1 WHERE major = 2 AND minor != next_beacon_minor ORDER BY phone, reading_timestamp; PGConf Silicon Valley @barrettclark
79. PGConf Silicon Valley @barrettclark
80. OTHER USEFUL THINGS PGConf Silicon Valley @barrettclark
81. POSTGRES • Install from source • Install via package manager • Heroku's Postgres.app • Postgres 9.6.1 • PostGIS 2.3.0 PGConf Silicon Valley @barrettclark
82. POSTGRES • Database GUI tools • pgAdmin3/pgAdmin4 (free) • Navicat (free trials) PGConf Silicon Valley @barrettclark
83. POSTGRES • Visual query plan • http://explain.depesz.com • http://tatiyants.com/pev PGConf Silicon Valley @barrettclark
84. http://explain.depesz.com PGConf Silicon Valley @barrettclark
85. http://tatiyants.com/pev PGConf Silicon Valley @barrettclark
86. CAUTION!!! Different environments are different PGConf Silicon Valley @barrettclark
87. DATATYPES • Array • DateRange and TSRange • JSON, JSONB (9.4) • UUID http://edgeguides.rubyonrails.org/active_record_postgresql.html PGConf Silicon Valley @barrettclark
88. UPSERT!!! New in Postgres 9.5 PGConf Silicon Valley @barrettclark
89. PARALLEL SEQUENTIAL SCAN, JOINS, AND AGGREGATES New in Postgres 9.6 PGConf Silicon Valley @barrettclark
90. FUTURE YOU WILL THANK YOU PGConf Silicon Valley @barrettclark
91. RECAP PGConf Silicon Valley @barrettclark
92. IT'S OK TO WRITE SQL PGConf Silicon Valley @barrettclark
93. REFACTOR • Common Table Expression (CTE) • View • Materialized View PGConf Silicon Valley @barrettclark
94. I'M NOT SAYING DO EVERYTHING IN THE DATABASE PGConf Silicon Valley @barrettclark
95. POSTGRESQL IS AWESOME PGConf Silicon Valley @barrettclark
96. PGConf Silicon Valley @barrettclark
97. THANK YOU! PGConf Silicon Valley @barrettclark