Skip to content

Instantly share code, notes, and snippets.

@sheki
Created April 24, 2012 06:24
Show Gist options
  • Select an option

  • Save sheki/2477094 to your computer and use it in GitHub Desktop.

Select an option

Save sheki/2477094 to your computer and use it in GitHub Desktop.
Cassandra Schema discussion
[11:11:57] <shekispeaks> are there any limits on the data size which can be stored against a super-column
[11:15:01] <driftx> http://www.quora.com/Cassandra-database/Why-is-it-bad-to-use-supercolumns-in-Cassandra don't use supers
[11:17:27] patricioe (~patricioe@99-28-161-222.lightspeed.miamfl.sbcglobal.net) joined the channel.
[11:18:18] Torgan (~Torgan@vil31-1-78-245-104-71.fbx.proxad.net) left IRC. (Quit: This computer has gone to sleep)
[11:18:51] mmalone (~mmalone@c-24-130-137-135.hsd1.ca.comcast.net) left IRC. (Read error: Connection reset by peer)
[11:19:24] mmalone (~mmalone@c-24-130-137-135.hsd1.ca.comcast.net) joined the channel.
[11:20:19] <nickmbailey> use composite columns instead: http://www.datastax.com/dev/blog/introduction-to-composite-columns-part-1
[11:20:42] <shekispeaks> ha thanks
[11:21:01] <nickmbailey> or just use regular columns
[11:21:35] wlian (~wlian@c-76-126-197-65.hsd1.ca.comcast.net) left IRC. (Quit: wlian)
[11:25:25] simsky (db850001@gateway/web/freenode/ip.219.133.0.1) joined the channel.
[11:25:40] mstump (~user@c-24-23-156-157.hsd1.ca.comcast.net) left IRC. (Ping timeout: 272 seconds)
[11:25:58] wlian (~wlian@c-76-126-197-65.hsd1.ca.comcast.net) joined the channel.
[11:26:40] simsky (db850001@gateway/web/freenode/ip.219.133.0.1) left IRC. (Client Quit)
[11:29:32] wlian (~wlian@c-76-126-197-65.hsd1.ca.comcast.net) left IRC. (Client Quit)
[11:29:56] detaer1 (~Adium@c-69-181-203-9.hsd1.ca.comcast.net) joined the channel.
[11:31:20] <shekispeaks> well looking at cassandra -> i have a use case where i have objects of different types -> the types are O(100s).
[11:31:29] <shekispeaks> i need to query give all objects of type t
[11:31:46] <shekispeaks> i was planning to use super columns for this. but now it looks like a bad idea
[11:31:54] <shekispeaks> what could be the best way to do this.
[11:32:53] detaer (~Adium@c-69-181-203-9.hsd1.ca.comcast.net) left IRC. (Ping timeout: 260 seconds)
[11:33:41] ik04ka (~igor@78.26.128.183) joined the channel.
[11:36:30] bondar (~dnjaramba@41.72.193.86) left IRC.
[11:37:34] graeme_f (~graeme_f@host109-158-98-22.range109-158.btcentralplus.com) left IRC. (Quit: Computer has gone to sleep.)
[11:37:42] Guest64169 (~ono_takes@27.131.11.254) left IRC. (Ping timeout: 246 seconds)
[11:37:55] <nickmbailey> type is your row key and each object is a column in that row, when you want to get all objects of type t you just get the entire row
[11:38:18] jijoe (~jijoe@99-59-181-145.lightspeed.sntcca.sbcglobal.net) joined the channel.
[11:39:01] davidstrauss (~straussd@wikimedia/davidstrauss) joined the channel.
[11:45:41] <shekispeaks> does this not load all the columns always?
[11:46:06] <shekispeaks> and what would be the limit on no. of columns i can store against a row.
[11:46:30] <driftx> two billion
[11:47:14] dpino (~dpino@159.20.76.188.dynamic.jazztel.es) joined the channel.
[11:47:16] <driftx> and you don't have to get all the columns at once per se, you can slice and paginate through them
[11:48:45] franco65 (~francois@93-50-77-16.ip151.fastwebnet.it) joined the channel.
[11:49:01] <shekispeaks> and how many column families can cassandra have -> any limits on that?
[11:49:24] franco65 is now known as ciao
[11:50:57] ciao (~francois@93-50-77-16.ip151.fastwebnet.it) left the channel.
[11:52:06] <driftx> thousands
[11:53:31] <shekispeaks> will it be a bad idea if my type is the column family
[11:53:40] <shekispeaks> and rows in it are the objects?
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment