Date: Thu, 13 Mar 2008 08:08:50 -0700
Reply-To: Irene <irenelj23@GMAIL.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Irene <irenelj23@GMAIL.COM>
Subject: Re: reading big size dataset
Content-Type: text/plain; charset=ISO-8859-1
On Mar 13, 5:38 am, gerhard.hellrie...@T-ONLINE.DE (Gerhard
> I don't understand, why you don't want to assign formats! The formats
> don't change anything of the size of your dataset. Formats are only in the
> header of the dataset, not with the data. That might be a trick to reduce
> (!) the size of a big dataset: if you have long text to explain certain
> information to be stored in the dataset, you might store short keys
> Example: you have a zip-code and a long name of city in your dataset. You
> could threw the city name out (normalization!) and assign a format which
> gives you a name for each zip-code.
> If you don't have a format or you are not sure, use the option nofmterr:
> options nofmterr;
> data newdate;
> set very.big;
> where zip=12345;
> On Wed, 12 Mar 2008 21:15:11 -0700, Properly <irain...@GMAIL.COM> wrote:
> >I want to make sure some questions.
> >1).For big dataset, because of the big size, we do not keep the format
> >for the variables when save the dataset, is this correct?
> >2).When I try to read the variables from a very big dataset (more than
> >10M), I get many messages like:
> >ERROR: Format ... not found or couldn't be loaded for variable .....
> >does this also because of the lost format? I can not understand why
> >without format, I can not read them....
> >3), how I can read this kind of big size dataset?
> >Thanks for your help!
Thanks so much for your help. But could you tell me why this message
has appear when I try to open the dataset?
ERROR: Format Name not found or couldn't be loaded for variable Name