genji-test

Genji database testing

Test results

Following obtained by running main.go in this directory.

Command output:

2020/11/02 12:31:13 SELECT * FROM users WHERE email = "joe@admin.com": documents found: 1, time: 108.26µs
2020/11/02 12:31:13 SELECT * FROM users WHERE firstname = "Joe": documents found: 1, time: 104.586876ms
2020/11/02 12:31:14 SELECT * FROM users WHERE email = "fred@admin.com": documents found: 100000, time: 762.689846ms
2020/11/02 12:31:14 SELECT * FROM users WHERE firstname = "Fred": documents found: 100000, time: 776.828214ms
2020/11/02 12:31:14 SELECT * FROM users WHERE email = "mary@admin.com": documents found: 1, time: 93.02µs
2020/11/02 12:31:14 SELECT * FROM users WHERE firstname = "Mary": documents found: 1, time: 104.352365ms
2020/11/02 12:31:15 SELECT * FROM users: documents found: 100002, time: 700.470693ms
2020/11/02 12:31:15 SELECT * FROM users WHERE id = 100001: documents found: 1, time: 85.627572ms
2020/11/02 12:31:15 All done :-)

Notes:

• User insert time (100 users)
  • older workstation with spinning HD: 57.9ms
  • newer workstation with SSD: 41.49ms
• DB size
  • 1000 records: 524288 (524 bytes/record)
  • 100,000 records: 41951232 (419 bytes/record)
• does inserting a lot of records with the same indexed value cause insert times to grow?
  • insert time for 100,000 records averages 33.73ms, so index does not appear to affect write time with large record sets.
• does index automatically improve query?
  • in our test, we index email, but not firstname
  • search on value in a table of 100,000 entries
    • search by email (indexed) takes: 116us
    • search by firstname (not indexed) takes: 90ms
    • indexed search is roughly an order of magnitude faster, but even a raw search is still blazing fast for 100,000 records -- seems this database would scale quite well.