What is the order of being added to PDV ?

[ResoluteCarbon]

Hi all!

Regarding the lesson 1 programming 2 (Data step processing: Compilation).

The theory is:

"To build the PDV, SAS passes through the DATA step sequentially, adding columns and their attributes. The SET statement in this program is listed first, so all of the columns from the storm_summary_small table are added to the PDV along with the required column attributes name, type, and length.

If there are any other statements that define new columns, those columns are also added to the PDV. In this code, the LENGTH statement defines the character column Ocean with a length of 8. StormLength is the last new column, and based on the arithmetic expression, it's defined as a numeric column with a length of 8."

The lecture note is:

what I want to focus is the column "Basin". I am wondering why the column Basin is at this position. Because SAS will go through all the codes sequentially. So, they will exclude the old Basin and add a new Basin column at the end of the yellow table in PDV when stopping by the data statement 

Basin=upcase(Basin);

Thank you!

[Tom · Accepted solution]

Huh?

This statement

Basin=upcase(Basin);

is not going to make a new variable.  It is going the change the value of the current variable.

The only way it would make a new variable would be if BASIN had not been seen before.

If you look at the first three lines of the DATA step and the picture showing the order of the variables you can KNOW that BASIN was defined because it appeared in the input dataset.

Do you see why you can know that?

Spoiler

Ocean is the first variable mentioned explicitly in the code.  Since it is the 6th variable in the photograph of the order that means the five before it must have come from PG2.STORM_SUMMARY_SMALL.