Clyatt: Robscow: Could be. Better load them one after another
Clyatt: Kittinger: Which means? :
Meekins: When do you actually use column oriented database ?
Salvaggio: Ndb is not webscale, right?
Willcox: Salle, aye that’s what i’m having to do, but it’s several hours every day
Clyatt: Robscow: Several hours every day? Something is wrong with your approach then
Clyatt: Robscow: Or the server is badly configured
Remis: Salle, hence my presence here :
Bourgault: Salle, Kahoun: column-oriented db stores tables as sections of columns of data rather than as rows of data.
Rill: So mysql is row oriented, and vertica is column oriented
Axon: Salle: but I suspect you know that
Quanstrum: Sections of columns of data. hum.
Lacina: Kahoun: it doesn’t appear that Vertica is in-memory,
Clyatt: Kahoun: MySQL Cluster is not exactly MySQL, but it is row oriented too
Clyatt: If “row oriented” really makes sense
Kaneta: Kahoun: btw there is Nelles-ndb for MySQL Cluster questions.
Clyatt: Kahoun: MySQL Cluster is shared-nothing cluster. I don’t see anything that suggests Vertica is a cluster like
Brosnan: Kahoun: it’s a more intimate affair over there.
Clyatt: Kahoun: It will help if you ask more specific question
Othon: I’m more interested in why would one use column oriented database and how to imagine it in my head : cause mysql table is simple excel spreadsheet if you imagine it.
Clyatt: Kahoun: You can’t be more wrong than that :
Clyatt: Kahoun: MySQL table has nothing in common with spreadsheet
Clyatt: Kahoun: s/MySQL/Any_RDBMS/
Boulay: Ye well you can export data in spreadsheet format so you can visualise data in the table like that.
Hasker: Kahoun: it’s better for sequential record access where you need to access the same field of multiple sequential rows rather than multiple fields of a single row.
Arnoldi: Kittinger, okay this is great answer. ok i get it
Clyatt: Kahoun: You can export JPEG file into array of bytes. Does it mean JPEG is mere array of bytes?
Clyatt: Kahoun: Not to mention with databases you export the result of a query not the actual data directly
Fuselier: Salle i totally understand that.
Clyatt: Kahoun: column oriented database refers more to the way the data is organized and has nothing in common with how it appears to the end user
Pean: Kittinger, does vertica make map/reduce like hadoop ? to spread load between nodes ?
Clyatt: Kahoun: Why are you asking that in Nelles channel?
Phalen: Cause there is no #vertica channel :
Tiogangco: And some of you may know it
Takenaka: I’m sorry if i’m off topic
Holzem: Kahoun: Back to Wikipedia. https://en.wikipedia.org/wiki/Vertica “The product integrates with Hadoop12 to leverage HDFS within Vertica and provide access to Vertica’s data through MapReduce.”
Leimkuhler: Kittinger, salle, thanks for the explanations.
Sayavong: Kahoun: if you are curious about MySQL Cluster. http://mysql.com/why-mysql/white-papers/mysql_cluster_eval_guide.php
Dzierzanowski: Can pattern matchingusing LIKE be based on 2 columns?
Bainbridge: Where col1 like %P%.that is for one column
Geesaman: Dimi1947: do you want fulltext ? MATCHa, b AGAINST ‘string’
Hyland: Dimi1947: what do you want to do?
Osden: Do you want me to describe the problem I am facing?
Nyce: I have 3 tables.by the way where to put the show create result?
Woehr: So you can see the structure of my tables
Quinnie: Anywhere except pastebin.com
Meaders: Dimi1947: http://pastie.org or http://sqlfiddle.com/
Lucente: Ok give me a minute or so to prepare it