Data Code Diary Entries

This is part of the Semicolon&Sons Code Diary - consisting of lessons learned on the job.

• Always use strings keys in JSON in languages that support symbol keys
• Be alert for non standard base representations
• Ensure all envs have same key seed accounts
• Format data before validating
• Has one and many2many etc relationships are about possibility not reality
• Higher DB ids should always correspond with later points in time
• If data is hashed or encrypted remember to check your setters do that too
• JSON requires double quotes for strings
• Prefer two boolean columns to one column which contains combinations
• Take data interchange specs seriously
• Use a key value built in schema with JSON
• Use fuzzy match algorithms to manage messy string data
• Use number types instead of stringifying everything
• Use simple data formats
• Use standard formats for date times
• Validations for presence are ambiguous with key value styles
• When append writing CSV do not write headers twice
• When modifying code that acts on production data test against a good sized sample
• When using foreign keys seed the dependent records first
• Your seed data must exactly reflect the structure of real data