turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- Data Management
- /
- Forum
- /
- how to fill up a dataset with samples till a certa...

Topic Options

- RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

01-04-2018 10:14 PM

Hi,

I have two data sets a and b with 200 entries each. While we have the capacity to test 350 samples each month, I have to use one of the data sets each month (in alternating mode so a b a b etc), and top the remaining places up with samples from the other dataset.

So say, for January, I use dataset a and top it up with 150 samples of dataset b.

In this case I could simply use Proc Sql... obs = 150

However, the samples set sizes change every month.

What do I need to do to :

choose a and fill up samples from sample set b till samples size = 350.

Any ideas are highly welcome.

Many thanks

Accepted Solutions

Solution

01-07-2018
03:16 PM

- Mark as New
- Bookmark
- Subscribe
- RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Posted in reply to Anna_NZ

01-04-2018 10:59 PM - edited 01-04-2018 11:02 PM

Here one way to go:

```
data have_a;
ds='A';
do i=1 to 200;
output;
end;
run;
data have_b;
ds='B';
do i=1 to 200;
output;
end;
run;
%let n_sample=350;
%let ds_all=have_a;
%let ds_sample=have_b;
data want(drop=_:);
/* ds with all records */
do while(not last_A);
set &ds_all nobs=nobs_A end=last_A;
output;
end;
/*** based on: http://support.sas.com/kb/24/722.html ***/
/* Method 3: Using SAS DATA Step with no sort required */
/* Initialize _K to the number of sample obs needed and _N to the */
/* total number of obs in the data set. */
_k= &n_sample - nobs_A;
_n=nobs_A;
do while(1);
set &ds_sample;
/* To randomly select the first observation for the sample, use the */
/* fact that each obs in the data set has an equal chance of being */
/* selected: k/n. If a random number between 0 and 1 is less than */
/* or equal to k/n, we select that the first obs for our sample */
/* and also adjust k and the number of obs needed to complete the */
/* sample. */
if ranuni(0) <= _k/_n then
do;
output;
_k=_k-1;
end;
/* At every iteration, adjust N, the number of obs left to */
/* sample from. */
_n=_n-1;
/* Once the desired number of sample points are taken, stop iterating */
if _k=0 then leave;
end;
stop;
run;
```

All Replies

Solution

01-07-2018
03:16 PM

- Mark as New
- Bookmark
- Subscribe
- RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Posted in reply to Anna_NZ

01-04-2018 10:59 PM - edited 01-04-2018 11:02 PM

Here one way to go:

```
data have_a;
ds='A';
do i=1 to 200;
output;
end;
run;
data have_b;
ds='B';
do i=1 to 200;
output;
end;
run;
%let n_sample=350;
%let ds_all=have_a;
%let ds_sample=have_b;
data want(drop=_:);
/* ds with all records */
do while(not last_A);
set &ds_all nobs=nobs_A end=last_A;
output;
end;
/*** based on: http://support.sas.com/kb/24/722.html ***/
/* Method 3: Using SAS DATA Step with no sort required */
/* Initialize _K to the number of sample obs needed and _N to the */
/* total number of obs in the data set. */
_k= &n_sample - nobs_A;
_n=nobs_A;
do while(1);
set &ds_sample;
/* To randomly select the first observation for the sample, use the */
/* fact that each obs in the data set has an equal chance of being */
/* selected: k/n. If a random number between 0 and 1 is less than */
/* or equal to k/n, we select that the first obs for our sample */
/* and also adjust k and the number of obs needed to complete the */
/* sample. */
if ranuni(0) <= _k/_n then
do;
output;
_k=_k-1;
end;
/* At every iteration, adjust N, the number of obs left to */
/* sample from. */
_n=_n-1;
/* Once the desired number of sample points are taken, stop iterating */
if _k=0 then leave;
end;
stop;
run;
```

- Mark as New
- Bookmark
- Subscribe
- RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Posted in reply to Anna_NZ

01-04-2018 11:58 PM

Grab the number of records in each table. Subtract them and create a macro variable with that value, then use that in your SQL code.

Here’s some ways to get those numbers

http://www.sascommunity.org/wiki/Determining_the_number_of_observations_in_a_SAS_data_set_efficientl...

Here’s some ways to get those numbers

http://www.sascommunity.org/wiki/Determining_the_number_of_observations_in_a_SAS_data_set_efficientl...