topic Re: Pulling required observations from 2 data sets in SAS Programming

Pulling required observations from 2 data sets

ysantosh18 — Fri, 24 Feb 2017 15:18:33 GMT

Hi,

Kindly help me, I want to pull required data from 2 dataset with a data step.

eg:

datasetA:

101 bike

102 car

103 bike

104 car

datasetB:

105 car

101 bike

102 car

I want to exclude the observations which are not in datasetB and in datasetA as follow

datasetC:

101 bike

102 car

103 bike

105 car

Note: here In C observation: 104 should be missed , and 105 should be there

Thanks in Advance

Santosh

Re: Pulling required observations from 2 data sets

art297 — Fri, 24 Feb 2017 15:28:36 GMT

Why is 103 bike in the desired set. Are you matching on one or both variables?

Re: Pulling required observations from 2 data sets

ysantosh18 — Fri, 24 Feb 2017 15:30:56 GMT

No, I want to keep the records of A and B and exclude the record which is not in B and in A

Re: Pulling required observations from 2 data sets

nehalsanghvi — Fri, 24 Feb 2017 15:34:03 GMT

103 bike is one of two records not in B and in A, so why have you put it in your desired output dataset C?

Re: Pulling required observations from 2 data sets

LinusH — Fri, 24 Feb 2017 15:34:21 GMT

Perhaps it's because of long working week, but I don't get the logic. Perhaps if you describe the business requirement it would be a bit clearer.
Also, why data step if e.g. SQL is better suited?

Re: Pulling required observations from 2 data sets

ysantosh18 — Fri, 24 Feb 2017 15:43:37 GMT

Am working with Clinical data, for Subject Elements, I need to set them One by One, if they are not in datasetB, they are said to be Non-completers of the study, so I've to exclude those subjects from A

thanks

Re: Pulling required observations from 2 data sets

nehalsanghvi — Fri, 24 Feb 2017 15:48:03 GMT

This will give you everyone from A who is also in B (i.e. excludes those from A who are not in B)

create table C as
select *
from A
where yourIDVar in (select yourIDVar from B);

Re: Pulling required observations from 2 data sets

ysantosh18 — Fri, 24 Feb 2017 15:55:56 GMT

thanks for, but it worked for records who are in both the datasets

thanks

Santosh

Re: Pulling required observations from 2 data sets

nehalsanghvi — Fri, 24 Feb 2017 16:02:04 GMT

Let's break this down Santosh. You do not want those obs from A which are not in B, This throws out:

103 bike

104 car

You say you do want 105 car in dataset C:

In this case, are you simply not looking for the dataset B itself? What are we not getting?

Re: Pulling required observations from 2 data sets

ysantosh18 — Fri, 24 Feb 2017 16:05:18 GMT

Yes!

Re: Pulling required observations from 2 data sets

nehalsanghvi — Fri, 24 Feb 2017 16:07:01 GMT

Okay, so are you all set? You just need dataset B? There is no coding problem to solve?

Re: Pulling required observations from 2 data sets

ysantosh18 — Fri, 24 Feb 2017 16:10:06 GMT

Not like that, As I said have to exclude the records which are "not in B" and "in A", Have to keep the records which are "in B" and "not in A"

Re: Pulling required observations from 2 data sets

art297 — Fri, 24 Feb 2017 16:12:35 GMT

Obviouslly, those of us reading this trhead don't understand what you are trying to do.

For each of the 4 records in data set c, separately, please explain why they were selected.

Re: Pulling required observations from 2 data sets

ysantosh18 — Fri, 24 Feb 2017 16:33:04 GMT

thanks all for your support!

I apologise for not being clear, My situation is like that!

Working with clinical data.

Let us take an example of Acadamic class,

class6 and class8

Class8 may have promoted students who are succeded, and class6 are striving for success, and if students not in class8 of class6 are said to be failures, and if there are individual students of both the classes are said to be succeeded.

still am not clear

No problem

thanks again

Re: Pulling required observations from 2 data sets

Yavuz — Fri, 24 Feb 2017 20:53:03 GMT

Your description is called as "left join".

proc sql;
create table dataset_c as
select dataset_a.*, dataset_b.*
from dataset_a LEFT OUTER
join dataset_b on dataset_a.student=table_b.student; quit;