Data Formats
- 1. How I Turned to the Dark Side. Formats of Data Transfer
- 2. What file types are there? The 4/5 most popular are: CSV, TSV XML JSON YAML
- 3. CSV / TSV Comma or Tab Separated Values Easy to dump into a spreadsheet Parsable using a library that uses SQL Little space CSV is not very human readable, and if large amounts TSV can get confusing
- 6. YAML YAML Aint Markup Language Very Human Readable Data structures Very useful for config and fixture files Easy for machines to read Whitespace dependent, so can produce very long files. Forced to be a particular structure
- 8. XML eXtensible Mark up Language If well-formed, can be read by a number of libraries Very common Whitespace independent Layout of data very much up to the individual - but needs documenting!
- 12. JSON JavaScript Object Notation Forced to be a particular structure Potentially Dangerous in JavaScript if just evalled Little memory space Data structure is obvious to a human reader if spaced out, although whitespace independent
- 15. What’s the best? Pros CSV/TSV good for sending to Spreadsheets and Databases YAML is great when it needs to be human modifiable, such as fixture data/config files XML is very versatile in how to markup data JSON is very compact and easily parsed into objects
- 16. What’s the best? Cons CSV/TSV can be difficult to work with in apps, as no variable names necessarily associated. YAML can be very long files, and needs to adhere to the whitespacing
- 17. What’s the best? Cons XML can be confusing if not well documented, and can be longwinded to obtain the information JSON can be less human readable if you are aiming for reducing bandwith by stripping whitespace
- 18. What is my Choice? Depends on the application, but: I want the data to be both Human and Machine readable I want the format to be well defined I want it to be convenient to parse I want it to be supported long term
- 19. What is my choice? Was XML, Fast becoming JSON Easy to parse Follows rigid structure If laid out it can be Easily eyeballed for the data Easily hand-modified