Is the answer correct in the book of Learning SAS® by Example

[jz127323]

This is a answer from Learning SAS® by Example: A Programmer's Guide, Second Edition. But I think the rusults are not the what expected to get. 

Q: Using the permanent SAS data set Survey1, create a new, temporary SAS data set (Survey1)
where the values of the variables Ques1–Ques5 are reversed as follows:
1  5; 2  4; 3  3; 4  2; 5  1.

*13-1;
data survey1;
set learn.survey1;
array Ques{5} $ Q1-Q5;
do i = 1 to 5;
Ques{i} = translate(Ques{i},'54321','12345');
end;
drop i;
run;
title "List of SURVEY1 (rescaled)";

proc print data=learn.survey1;
run;
proc print data=survey1;
run;

List of SURVEY1 (rescaled)

 Obs Subj Q1 Q2 Q3 Q4 Q512345

001	1	3	5	4	2
002	5	5	4	4	3
003	2	1	2	1	1
004	3	5	1	4	2
005	3	3	3	3	3

---

List of SURVEY1 (rescaled)

 Obs Subj Q1 Q2 Q3 Q4 Q512345

001	5	3	1	2	4
002	1	1	2	2	3
003	4	5	4	5	5
004	3	1	5	2	4
005	3	3	3	3	3

[ballardw]

Where specifically do you have a question? Which specific values? Which observation (row)?

We don't have that data set so you need to be very explicit about your concern. From what I see the second output has all of the 1 changed to 5 and vice versa and 2 changed to 4 and vice versa. Isn't that what the introduction says to do?

(Not the way I would do this and actually not needed at all for 99% of activitiy, but these are learning examples)

You really need to provide two different TITLES for the two proc prints because your title implies that they are the same output but it isn't. One of the titles should say something about "before transformation" and the other "after transformation".

[jz127323 · Accepted solution]

The first table is the original dataset

the second is after transformation

[Reeza]

Results look correct to me.

[jz127323]

the old obs1: 

001

1

3

5

4

2

the new dataset, should it be:

001

2

4

5

3

1

thanks

[andreas_lds]

---

@jz127323 wrote:

the old obs1: 

001

1

3

5

4

2

the new dataset, should it be:

001

2

4

5

3

1

thanks

---

Where is the problem? The new dataset is correct.

Note: I compared the original datasets, not the obs posted here.

@jz127323  you may want to show the code you have written, as text using "insert sas code" button, so that we can actually help.

[Kurt_Bremser]

---

@jz127323 wrote:

the old obs1: 

001

1

3

5

4

2

the new dataset, should it be:

001

2

4

5

3

1

thanks

---

No. From your description, 1 changes to 5 and 2 to 4 (and vice versa), while 3 remains unchanged. 

[Reeza]

---

@jz127323 wrote:

the old obs1: 

001

1

3

5

4

2

the new dataset, should it be:

001

2

4

5

3

1

thanks

---

No, reverse coding means reverse the values individually so that :

5 -> 1

4 -> 2

3 -> 3

2 -> 4

1 -> 5

You seem to want to reverse the order of the values in the row, ie what is in col5 goes to col1 which is not the correct interpretation of the question. 

[jz127323]

Thanks for everyone's help. I totally misunderstood the meaning of the question. Not to exchange the values of columns, just change the value itselves. 

Thanks again!