adding a configuration type (masking.yml) that doesn't change data cardinality

[kibitan]

background

when we want to use anonymized data for performance tuning purpose, we don't want to change data cardinality in records, otherwise, it is not showing the correct sql execution plan.

grammar

pass with digit

users:
  email:  anonymized+%{hash,4}@example.com

expected result

before

expected result

MySQL [mydb]> SELECT id, email FROM users ORDER BY id;
+----+----------------------------+
| id | email                      |
+----+----------------------------+
|  1 | test1@gmail.com            |
|  2 | some1@hotmail.com          |
|  3 | test1@gmail.com            |
+----+----------------------------+

MySQL [mydb]> SELECT id, email FROM users ORDER BY id;
+----+----------------------------+
| id | email                      |
+----+----------------------------+
|  1 | anonymized+fdas@example.com | <- same data outputs same data
|  2 | anonymized+32bf@example.com |
|  3 | anonymized+fdas@example.com | <- same data outputs same data
+----+----------------------------+

Tech tips

• it can be a performance regression

[kibitan]

this is relates with #71

[kibitan]

check this repo too: snaplet/copycat and this: docs.snaplet.dev/references/data-operations/transform

[kibitan]

@tonklon voted for this feature in 2014 Sep at frankfurt.rb, thank you for the feedback!