谢谢上传

63题:
Q2
Q3
Q4 - 变形，考了两道变形：
（1）        题目给了C选项，问what groups will be stored?  选0-29，30-49，50+
（2）        通过SQL中select into：clause生成新的macro变量。
Desired output: “&Shoesales1.:&Shoestype1.,
&Shoesales2.:&Shoestype2.,
&Shoesales3.:&Shoestype3.,
”
选INTO :Shoesales1- :Shoesales3, :Shoestype1- :Shoestype3
注意：a.冒号的位置  b.中间的横杠
干扰项：INTO: Shoesales1- Shoesales3, :Shoestype1- Shoestype3
干扰项的shoesales3和shoestype3前没有冒号，注意

Q9 - 变形 变成了
Jones 105
Jones 105
Smith 105
Smith 105
Smith 105
选A没有group by ，min（cost+ship）的那个
Q10 -变形 换成了11和5 选2

Q11 - 答案不变， 但是有改写 把 when 。。。。 dataset created。

Q13 - 变体；题目贴在后面MZX最强机经的Q13变体里。
Q14 - 答案不变，注意有个干扰项：describe table= xxx，注意不能有等号。
Q15- 变形，数字好像换了，换成sum(cost)了，还用了group by， 问output，就是按每个
人分别求和。
选  Jones  20
                  Smith 120
Q16 - 变形，desired：
Lauren L
Patel    A
选intersect corr.
Q18
Q19
Q21
Q23
Q25 – 变体No和no改变位置，Belt = no. 选2
Q27
Q29
Q30 – 换单词，把product变成book. 答案不变
Q37 – 变形，给了code用的except all，问output？
选 output :
Chang
Chang
Lauren       
Q38 – 变形，4和10换位置，选result = 0
Q39 - 变形bicycles 和shoes 换了位置, 选Mv is bicycles
Q40
Q41 – 变形，给具体log来判断，还是选择mlogic
        Log 大概长这样
xxxx(PRODUCT):  Beginning execution.
xxxx(PRODUCT):  Parameter MV has value tents
Q43 – 变形；变成desire Jones 100， 选项选having　avg(cost) < ()
Q44- 变形：
（1）        题目说要select table with columns name ended “Date” in the library “ORION”，选项差不多，选 dictionary.columns。
（2）        Sas 官网给的Sample 题，里面第5题，考的原题。
Q45 – 变形：3个选项用具体code给出的，选项的顺序变了，不选first.customer_id，因为没有sort。选剩下那俩。
Q47 - 单词换成Monday了，其他不变。
Q48 - 把want改成别的词了，其他不变。
        Where HireDate = “&Date”d
Q49
Q53，变形，多选题，题目中有提示说选2个。选的reduce cpu, reduce I/O (没有increase memory的选项)。
Q54
Q55，变形，desired的output里面删掉了year 2012那行，选 left join
Q58，题目类似，就选 is not missing
Q60，变形，题目中means和print互换位置，选proc print only。
Q62，--》变成了 ！！！，inside换成in，outside换成out，其他不变。
Q63，变形，给code var weight height %cols1; 问输出什么变量？选weight height name age （顺序要和题中给的一样，这个有干扰项是height在weight之前）

3道填空：
(1). ASK: SAS log 显示三个global macro variables，如何在Log 中輸出global macro variables
ANS: %put_GLOBAL_
FIB: _GLOBAL
(2) ASK: outer union corr让你再填个 keyword. 能和 outer union 一起用的
FIB:corr
(3) Given 2 Data Set
ONE
Name Year
Joyce         9
John         4
John         2
Jane         6
Thomas         8

TWO
Name   Age(不需要理會)
Joyce        
John
Robert
Jeff

The following SAS program is submitted:
select Name, avg(Year) as average
from WORK.ONE
Except Corr
WORK.TWO
…………..
quit;
(6+8) / 2 = 7
ANS: Average of  7
FIB: 7
官方SAMPLE 9道题
Q2原题；正确选项是Coalesce + ON
Q5原题；正确选项是NOBS + dictionary.tables

SAS ADVANCED 質素機經 by mikeleung110
新題庫updated by 02.Jan.2018 by MZX

ASK: A Data has 2000million observations and 300 Character variables
ASK:Compress=YES
FIB:YES
OR 選擇題
A COMPRESS=YES
B COMPRESS=HEX
C COMPRESS=BIN
D COMPRESS=NUM
ANS:A

Q:FCMP 选择题
proc fcmp outlib=sasuser.funcs.trial;
…
endsub;
options ____cmplib_____=sasuser.funcs;
data null;
…
run;

ANS: cmplib

选择题原题
Q:What option force the use of a specific index from dataset?
Ans: IDXNAME


Q:MACRO选择题
%let this_year=%substr(&sysdate9, 6);
%let next_year=&this_year+1;
%let check_year=%eval(&next_year<2016);
%put two years after this year is &next_year+1;
%put check_year is &check_year;
Assume system time is 01Jan2013, what is the output?

ANS: Two years after this year is 2013+1+1
check_year is 1

Q:Multi-array选择题(记住output)
Array multi{1:2, 2}(1,2);
Do i=1 to 2;
Do j=1 to 2;
Output=multi{I,j};
What are the corresponding values of i, j, and output.

A 2*2 multi-array, only two initial values, so the array is following

i  j output
1 1 1
1 2 2
2 1 .
2 2 .

Q:Pagesize info
Which of proc can check the pagesize info?
A. Proc Contents
B. Proc print
C. Proc report
D. Proc catalog

ANS:A

变体：选项改变了表达方式
Q: Pagesize Info
A. Content procedure
B. Catalog procedure
C. Print procedure
D. Report procedure


Q: KEEP&DROP statement and option Efficiency
A data set has 300,000 observations, 20 character variables, 50 numeric variables. We need 5 character
variables and 7 numeric variables, which one is the most efficient:
A. Drop=option in data step
B. KEEP=option in data step
C. Keep =option in set statement
D. Keep statement

Answer: Keep=option in set statement. With it the system only read the desired variables.



Q45 变形 选择题
(i)
Given data sets and SQL code with Select Distinct statement, which one has the same output.
The correct answer is proc SORT with Nodupkey, and with an option(drop=XXX).
Besides, BY variable cannot have the descending option, because in the SQL procedure, ORDER BY
statement is in ascending order. So choose the answer like following:

proc sort data=XXX
out=work.sorted (drop=xxx) nodupkey;
by order_var;
run;

IF First.Customer_ID choice doesn’t sort the data first, so cannot be used.

这道题的原题模式我尽量记下来了

WORK.SALES

ClientID        Name        Sales
400        xxx        00000
100        Xxx        …
200        Xxx        …
100        Xxx        …
300        Xxx        …
400        Xxx        …
100
……        Xxx        …

SAS program is submitted: (应该没记错...)
proc sql;
select distinct
ClientID
Name
from Sales;
order by ClientID;
run;

Output:
ClientID        Name
100        Xxx
200        Xxx
300        Xxx
400        xxx

A. proc sort data = Sales  out = work.sorted (drop Sales);
        by ClientID;
     run;
B. proc sort data = Sales  out = work.sorted nodupkey;
        by ClientID;
     run;
C. proc sort data = Sales  out = work.sorted (drop Sales) nodupkey;
        by ClientID;
     run;
D. proc sort data = Sales  out = work.sorted (drop Sales) nodupkey;
        by descending ClientID;
     run;

我选C. 感觉机经里面也是这个意思。
正确答案具备以下特点：
1.        有(drop Sales)，可以看到output里已经没有Sales
2.        有nodupkey, 符合题目给的SAS code - “proc sql; select distinct…”，创造unique key
3.        给的code和output都能看出是按照ascending ClientID排列。D肯定错。


Q:Using View
Data finance .newdata/view=finance.newdata;
Infile<fileref>;
<Data step statements>;
run;

Submit the above code and create a data step view, then we need to use this view in the PROC MEANS
procedure, which one to use:

A. Proc Means view=finance.newdata;
B. Proc Means data= finance.newdata/view= finance.newdata;
C. Proc means data finance.newdata/view
D. Proc means data= finance.newdata

Answer: D. proc means data=dataset_name. The same way as to use a normal data set.
选没有view最短的那项

Q:MACRO
Given data set and macro program, choose missing correct code
A. &Num=California
B. call symputx(‘Num’, California)
有个干扰项是symputx(‘Num’, California)，别漏了call
Answer: B

Q: left join and in-line view
Product
Product_id Product
1                1001
2                1002
3                1003
Sales
Product_id Sales
3                100
1                200
5                100
1                200
3                100
1                100

Proc sql;
select p.product, s.totalsales
from product as p
left join
(select product_id, sum(sales) as totalsales
from sales
group by product_id) as s
on p.product_id=s.product_id;
quit;
What is the output?
Answer:
Product Totalsales
1001 500
1002 .
1003 200


Q: in-line view
给了一段code 明确告知in-line view 中给定的condition 有
multiply observations satisfied the condition， 问program 运行结果。

Q：
答案是运行出错没有结果，因为in-line view return multiple results.
Ask why program failed,
答案order by statement can't be used inside of an in-line view;
注意干扰项：order by cannot be used in a query where there is an in-line view.


Q: When would it be most efficient to use multiple SET statements in the DATA Step?

A.        One file has large amount of rows of data, another one only has small number of data
Index is used as key (这句表述我不太确定) and大的dataset 里有index

B.        Both files have large amount of rows of data
Index is used as key

C.        One file has large amount of rows of data, another one only has small number of data
Index is NOT used as key

D.        Both files have large amount of rows of data
Index is NOT used as key

zwl199464说：我觉得这道题是A 主要是根据set 和 key option 在书中的描述 不过不是很确定 望指正
MZX选了A 不确定。
我也选A了


Q:没给出具体数据，但说了关于这两个数据集的一些限定条件。要选出inactive_date 为missing 时候的observation。题目是
prol sql;
Select *
From dataset1 A
Full join dataset2 B
On A.ID = B.ID
Where inactive_date is missing;
问 data dataset3;
merge dataset1 (In=Ina)
dataset2 (In=Inb);
By ID;
<insert statement>;
选项有 A: where inactive_date is missing and (Ina and Inb)
B: where inactive_date is missing and (Ina or Inb)
C: if inactive_date is missing and (ina and inb)
D: If inactive_date ne .

MZX选择C 不确定
XM  (07May2019)： 我选了B，就是or那个。因为用的是full join.
我也选的B，题目是full join
感觉这题是有变体的，大家要看好full join还是merge或者其他join方式


Q： 考in line view
proc sql;
create table forecast as
select a.*, b.sales
from actual a, _________
where a.dept=b.dept;
quit;
ANS:
(select avg(revenue) as average
from Budegt
group by 1)  b
MZX注释：考的是括号的位置，记清即可
XM  (07May2019)：里面有干扰项，
select (avg(revenue) as average
from Budegt
group by 1 )  b
还有俩干扰项用的create，比如
(create avg(revenue) as average
from Budegt
group by 1)  b
有内容在括号外的可以快速排除，剩下两个（select...）(create…),选 select那个


Version1：什么样的index can only be set up with DATASET procedure.
A: index centiles can be updated.
B: 可以有mutiple index ？
C: index只有numeric ？
D:index 可以包括missing value
MZX选A 不确定
我也选A

我考的原题如下：
Q：What characteristic of index can only be set by creating index with DATASET process?
A. When index centile upgraded
B. The variable of index is required to be unique
C. Index include many variables
D. Index allows to have missing value.



Verision 1: 用where statement 怎样的index 最有效？我选择了index that disqualify most variables
Verision 2:一个observation 很多的dataset,。问选什么index 最好
我比较犹豫的两个选项是
A. XXXX
B. index qualify most variabes
C: index can disqualify most variables.
D: a composite index
MZX选C 不确定
我选了C

我考的原题如下：
Q：When searching in a large dataset with multiple indexes using Where statement, which is the best index to use?
A. an index disqualify as many as observations
B. an index includes as many as observations
C. composite index
D. index with a numeric variable
我个人认为composite index不一定是最好的index，关键在于放在一起作composite的variable/column能不能做到unique。
比较好的composite index例子，姓名+生日
比较不好的composite index 例子，姓名+性别（同名同性别的人可能很多啊）
所以按这个思路A比较可行。



SAS最强机经 by MZX

Q13 变形

我考试的时候是WORK.ONE有unsorted ID, 有时是WORK.TWO但道理一样。
原来的dataset Unsorted, Output里面的ID都被sort了。
正确答案：有FULL JOIN+coalesce


質素机经也提到这题：是同一道题
Q:给了两组data并给出了output，让选择能生成同样output的code.注意到数据one没有sort 而output中ID是sort过的。
A. Data Set MERGE; by Var；
B 和 C都是明显错
D.PROC Sql 的FULL JOIN+coalesce 但没有order by
ANS：MZX选D（PS.我不知道但没有order by是什么意思，考试选项如上，是有order by的）




我考的原题：针对proc sort问题里的（2）
Q：给一个dataset + 一个output问怎样generate same output?
Dataset里面包含多个相同Customer_ID
Customer_ID        Name        Sales
100        aaa       
100        aaa       
200        bbb       
200        bbb       
500        ccc       
               

Output
Customer_ID        Name        Sales
100        aaa       
100        aaa       
200        bbb       

A. data xxx;
set xxx;
by customer_ID;
if last.customer_id then output;
B. proc sort data=xxx out=xxx Nodup;
by customer_ID;
C. proc sort data=xxx out=xxx Nodupkey;
by customer_ID;
D. data xxx;
set xxx;
order by customer_ID;(没有group by)
答案选C
这道题不确定D具体是怎么写的，反正是很普通的create dataset。只记得好像没有group by
判断的方法：last.var肯定错了，剩下两个proc sort其中一个的nodupkey写错了，那就选了剩下那个。

前面提到的Q13变形，出自SAS最强机经 by MZXSAS最强机经 by MZX
Q13 变形

我考试的时候是WORK.ONE有unsorted ID, 有时是WORK.TWO但道理一样。
原来的dataset Unsorted, Output里面的ID都被sort了。
正确答案：有FULL JOIN+coalesce

图po在这里