topic variable size of data sets in SAS Procedures

variable size of data sets

bk53 — Thu, 11 Apr 2019 09:31:31 GMT

Hi,

is there a way to change the physical size of outputfiles?

I want to write a macro, which create testfiles with a determined size, the content does not matter.

Thanks

Re: variable size of data sets

Jagadishkatam — Thu, 11 Apr 2019 10:32:35 GMT

Could you please give more details, you want to know the size the datasets or change the size of the datasets

Re: variable size of data sets

bk53 — Thu, 11 Apr 2019 10:37:50 GMT

I want to Change the size of the datasets.

My main Problem is, that i want to create a macro witth the Parameter size and this macro create testfiles wich have this size, but i cant find a way to do this.

thanks

Re: variable size of data sets

andreas_lds — Thu, 11 Apr 2019 10:52:57 GMT

@bk53 wrote:

I want to Change the size of the datasets.

My main Problem is, that i want to create a macro witth the Parameter size and this macro create testfiles wich have this size, but i cant find a way to do this.

thanks

bk

Size depends on the variables you create, the number of observations in the dataset and some other things like compression, page-size ... I should be possible to calculate the number of observations and variables to get a defined size. But i can't see a use-case requiring such files.

Re: variable size of data sets

bk53 — Thu, 11 Apr 2019 11:00:25 GMT

Thank you for your reply.

I need the data to test Transfer between to different Systems.

Is there a formula to calculate the size?

thanks

Re: variable size of data sets

Jagadishkatam — Thu, 11 Apr 2019 11:04:33 GMT

If it is about determining the size of the existing dataset in a library, we can know it by the below code.

The dataset sized will create a dataset where it will have name and dataset in library and size with other variables.

proc sql;
create table sized as select * from dictionary.tables where libname='name';
quit;

Re: variable size of data sets

bk53 — Thu, 11 Apr 2019 11:34:12 GMT

Thanks.

Here is an example what i need: a 3MB file, a 1GB file, a 912KB file.

I just want to create them, but the Content doesn't metter.

Thanks

Re: variable size of data sets

Astounding — Thu, 11 Apr 2019 12:31:44 GMT

You could always use a little trial and error. For example:

data test1;
retain var1-var99 0;
do var100=0 to &nrecs;
output;
end;
run;

After a few tests you'll have a pretty good idea of how large &nrecs should be.

Re: variable size of data sets

bk53 — Thu, 11 Apr 2019 13:10:55 GMT

I used for nrecs values from 1 to 1000 and the size is every time 192kb.

Did i have to Change some Options?

Re: variable size of data sets

Astounding — Thu, 11 Apr 2019 13:23:21 GMT

These values are way too small. 1 means create a data set with 2 records. Start with a million and go from there.

Re: variable size of data sets

ballardw — Thu, 11 Apr 2019 14:46:55 GMT

@bk53 wrote:

Thank you for your reply.

I need the data to test Transfer between to different Systems.

Is there a formula to calculate the size?

thanks

bk

First thing what are you using to measure "size"? Bytes, number of rows, number of characters, number of columns?

What type of file does this transfer need to use? Fixed column text files the size can be determined fairly easily but is dependent on how many characters per line.

Delimited text files such as CSV or Tab delimited would likely be not too difficult to create a fixed size but again content per line is important.

Binary files such as XLSX or structured files like XLM not so much. The XLSX can have a lot of overhead style information and the text format of XLM repeating variable names and types means that the actual names of variables affects total file size.

Re: variable size of data sets

andreas_lds — Fri, 12 Apr 2019 05:36:27 GMT

@bk53 wrote:

Thank you for your reply.

I need the data to test Transfer between to different Systems.

Is there a formula to calculate the size?

thanks

bk

Have you seen http://support.sas.com/kb/51/659.html?