 Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.

# Re: st: matrix colnames, Kronecker direct products, and factor variables

 From Steve Samuels To statalist@hsphsun2.harvard.edu Subject Re: st: matrix colnames, Kronecker direct products, and factor variables Date Sun, 8 Dec 2013 13:49:29 -0500

```Here's code that illustrates the problem again and gives a Mata solution

Steve

Steven J. Samuels
18 Cantine's Island
Saugerties NY USA
Voice: 845-246-0774

*************CODE BEGINS********************
matrix A = (1,2)
matrix B = I(3)
matrix colnames A = 1.x 2bn.x
matrix rownames A =  z
matrix colnames B = 1.y 2.y 3.y
matrix rownames B = b1 b2 b3
matrix C = A#B
matrix list C

mata:
/* Get crs: Row Stripe for C */
ars = st_matrixrowstripe("A")
brs= st_matrixrowstripe("B")
arsn =  rows(ars)
brsn = rows(brs)
crsn = arsn*brsn
crs = J(crsn,2,"")

for (j = 1; j <= crsn; j++){
crs[j,1] = ars[ceil(j/brsn),2]
crs[j,2] = brs[mod(j-1,brsn) +1, 2]
}

/*Get ccs: Column Stripe for C */
acs = st_matrixcolstripe("A")
bcs= st_matrixcolstripe("B")
acsn =  rows(acs)
bcsn =  rows(bcs)
ccsn = acsn*bcsn
ccs = J(ccsn,2,"")

for (j = 1; j <= ccsn; j++){
ccs[j,1] = acs[ceil(j/bcsn),2]
ccs[j,2] = bcs[mod(j-1,bcsn) +1, 2]
}

st_matrixrowstripe("C",crs)
st_matrixcolstripe("C",ccs)
end

matrix list C
************CODE ENDS********************

On Dec 4, 2013, at 3:02 PM, Doug Hemken wrote:

I don't get what I expect. The documentation I find says the columns of the product matrix, A # B, will have names from matrix B and equation names from matrix A?

But if I have factor variable names for the columns of A, they simply disappear in the product. (I also notice that when I assign column names that appear to be factor variables or interaction terms, Stata alters my labels, usually for the better.)

For example:
matrix A = (1,21 \ 0 ,1)
matrix colnames A = _cons mpg
matrix B = I(2)
matrix colnames B = 0.foreign 1.foreign

matrix C = B' # A
matrix list A
matrix list B
matrix list C

The column labeling on C is not what I expect!
--
Doug Hemken
4226I Social Science Bldg.

dehemken@wisc.edu
262-4327

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/
```