All of Oracle's (non-float) number types are variable size numbers with an ordinal and a mantissa. This makes Oracle number very efficient for smaller values as compared to fixed size integers, but less efficient with larger values. NUMBER has a maximum precision of 38 digits with a scale of -84 to +127. NUMBER consumes between 1 and 22 bytes on disk. It is typical to specify a NUMBER with (p, s). In the absence of definition, precision of 38 and scale indeterminate will be assumed.
The exception to this are IEEE floating point number types which are a fixed size regardless of value. Summary: Oracle has no fixed length equivlents to tinyint, smallint, int or bigint from other databases and can either store these values more or less efficiently than those databases with fixed length integer types. -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Scott Marlowe Sent: Friday, March 09, 2007 1:58 PM To: Devrim GÜNDÜZ Cc: Shane Ambler; Kevin Hunter; [EMAIL PROTECTED]; PostgreSQL General List Subject: Re: [GENERAL] "oracle to postgresql" conversion On Thu, 2007-03-08 at 16:05, Devrim GÜNDÜZ wrote: > Hi, > > On Fri, 2007-03-09 at 05:21 +1030, Shane Ambler wrote: > > NUMBER is Oracle's version of NUMERIC - Oracle will use both but > > probably only Oracle will use NUMBER. > > Really? I thought Oracle's NUMBER ~ PostgreSQL's (BIG)INT? Not sure. It let me assign a precision to it, so I figured it wasn't int based. In fact, it accepts precision up to 38, just like numeric, and it accepts non-decimal portions, i.e.: number(20,4); ---------------------------(end of broadcast)--------------------------- TIP 1: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly ---------------------------(end of broadcast)--------------------------- TIP 2: Don't 'kill -9' the postmaster
