· What are some good SAS programming practices for
processing very large data sets?
Answers were Sorted based on User's Feedback
Answer / sibin
Use
1.WHERE Condition
2.DROP KEEP in Set Statement
3.ARRAY
4.Mutually Exclusive Conditions like DO ,DO UNTIL,DO WHILE
| Is This Answer Correct ? | 16 Yes | 1 No |
Answer / pricil kurian
1) Compressing dataset
2) Using where instead of if as subsetting conditions
3) Deleting the temporary datasets at the end of the program
4) If we use "if..else ..if conditions"..use most occuring
situations first
5)Use keep and drop appropriately to avoid the unneceesary
variable
6).Creattion of index variable
7)Use of array statement
8)Create one byte required numeric variable which we don't
use for calcualting process as charcater..Using this this
variable need 1 byte where numeric variable's minimum
length is 3.
9)Minimize the CPU time that SAS uses by executing the
minimum number of statements in the most efficient
order.
10)Use SQL for code simplification. A single SQL statement
is generally easier to code and understand than a
long series of SORT and DATA steps.
11)Use of data _null_; when we don't need to output into a
new dataset.
12)Use macros for redundant code.
13)When data sets are large and unsorted, the SQL inner
join may beter than SORT and MERGE.
| Is This Answer Correct ? | 8 Yes | 0 No |
Answer / vicky
1.we can compress the dataset
2.if we wana do merge a bigdate set with another dataset
having few variables for instance a fact with dimension
having key variable and description column
in this case sorting n merging this bigdateset(fact) with
dimension apply formats.so dat u can avoid sort n merge.
| Is This Answer Correct ? | 2 Yes | 1 No |
Answer / dvsr
1)compress the datasets
2)use DROP KEEP options while creating temp data sets
3)Index creation on the Frequently using variables
| Is This Answer Correct ? | 0 Yes | 1 No |
One way of creating a new variable in Macros is by % Let....What is the other way..?
How necessary is it to be creative in your work?
what do you mean by data staging area? : Sas-di
describe about metadata object? : Sas-di
How could you generate test data with no input data?
Explain proc univariate?
What areas of SAS are you most interested in?
Do you prefer Proc Report or Proc Tabulate? Why?
If you need the value of a variable rather than the variable itself what would you use to load the value to a macro variable?
Do we follow ADAM in analysis dataset development?How? Usually which version? Why is it necessary?
what is AE onset date and what is RDS?
Which is the best training Institute for SAS Business Intelligence course?