How many bytes of memory is a tweet?

Question

140 characters. How much memory will it take?

I am trying to calculate how many tweets my EC2 Large Mongo DB instance can contain.

Answer 1

Twitter uses UTF-8 encoded messages .

UTF-8 code points can be six in length, four octets in length , which makes the maximum message size 140 x 4 = 560 8-bit bytes .

This, of course, is only for raw messages, excluding the costs of storage, indexing and other storage additions.

e: Twitter successfully allowed me to post:

™ need to be, depending on what you need, depending on what you need, depending on what you need, depending on what you need, depending on your choice ™ need to be, depending on what you need, depending on what you need, depending on what you need, depending on what you need, depending on your choice ™ needs, depending on what you are, depending on whether you are

Yes, these are 140 trademark characters, which are three octets in UTF-8

Answer 2

Back in September, an engineer on Twitter gave a presentation that suggested about 200 bytes per tweet.

Of course, you still have to consider the overhead for your own metadata and the database itself, but 200 bytes / record is probably a good place to start.

Answer 3

This is usually two bytes per character, if you save Unicode as UTF-8, so that would mean a maximum of 280 bytes per tweet.

Answer 4

Probably 284 bytes in memory (prefix of length 4 bytes + length * 2). I can’t say inside the database, but probably 280, if the database is UTF-8, you can add a few bytes of service data, metadata, etc.

Answer 5

Potentially interesting:
http://mehack.com/map-of-a-twitter-status-object
Twitter Status Object Anatomy

Also more about Twitter character encoding:
http://dev.twitter.com/pages/counting_characters

Answer 6

It is technically stored as UTF-8, and actually the slide deck from the guy on Twitter here http://www.slideshare.net/raffikrikorian/twitter-by-the-numbers gives a real stat about this:

140 characters, ~ 200 bytes