Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

Topic Options

- RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

12-13-2011 09:45 AM

Since many vectors are n x 1 upon definition in mathematical notation, while IML vectors are 1 x n upon creation, if I want to use a vector as n x 1, I have to transpose it or “j” it into a vertical rather than horizontal orientation. Is there an IML option that I do not know about that would automatically create vectors as n x 1?

- Mark as New
- Bookmark
- Subscribe
- RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Posted in reply to Tenno

12-13-2011 11:21 AM

There is no option. In statistics vectors appear in both directions. For example, in a regression context you might encounter Y as a coliumn vector of responses and X[,i] as the i_th explanatory variable. However, to score the i_th observation requires evaluating a model on X[i,], the i_th row.

I don't understand your comment that "IML vectors are 1 x n upon creation." Are you thinking of a particular function or operation? In SAS/IML, some functions and operations result in row vectors, such as the LOC and UNIQUE functions and set operations such as SETDIF and XSECT. Also, some functions such as MEAN, VAR, and STD naturally return row vectors. However, if you read data from a SAS data set or define x=j(10,1), then you get a column vector.

You can use the COLVEC function to ensure that a vector is a column vector. The SHAPE and SHAPECOL functions might also be useful.