need help on on colnumb or rownumb

[hong chang]

Any idea what's going with matric D1 and F1 in below. I have a matric D1 which contains some post estimation of from a margins command. But when I tried to get the column number of 1._at#1.Treatment in D1, the STATA issues a error message. But if I go one step further to let another mat F1 to be the same as D1 (mat F1=D1), it won't have problem with F1 for the same operation. What is going on?




. mat list D1

D1[9,6]
1._at# 1._at# 1._at# 2._at# 2._at# 2._at#
1.Treatment 2.Treatment 3.Treatment 1.Treatment 2.Treatment 3.Treatment
b .97819882 1.1457846 1.0719707 .75463889 .96552951 .88649156
se .01749367 .03349974 .03991627 .01724896 .03304899 .03991627
z 55.917298 34.20279 26.855479 43.749819 29.215101 22.208775
pvalue 0 2.20e-256 7.28e-159 0 1.25e-187 2.83e-109
ll .94391186 1.0801263 .99373623 .72083155 .90075468 .8082571
ul 1.0124858 1.2114429 1.1502052 .78844623 1.0303043 .96472602
df . . . . . .
crit 1.959964 1.959964 1.959964 1.959964 1.959964 1.959964
eform 0 0 0 0 0 0

. global aa=colnumb(D1, "1._at#1.Treatment")
type mismatch
r(109);

.
. mat F1=D1

. global aa=colnumb(F1, "1._at#1.Treatment")

. di "$aa"
1

[Paul Burkander]

Did you ever solve this? I'm having a similar issue where colnumb returns "type mismatch" but if I just assign that matrix to another matrix it works. So weird!

[Andrew Musau]

What version of Stata do you have? From version 16, the matrix functions -colnumb()- and -rownumb()- are superseded by more straightforward ways to refer to matrix columns and rows. See

help matrix extraction

for details. In any case, #1 most likely did not get a response due to lack of a reproducible example. You will increase your chances of making progress by providing one (FAQ Advice #12 explains what is required).

[William Lisowski]

The problem in post #1 is likely the result of having not only a Stata matrix named D1 but also a variable in the current dataset named D1.

. mat D1 = 1,2,3 \ 4,5,6

. mat list D1

D1[2,3]
    c1  c2  c3
r1   1   2   3
r2   4   5   6

. display colnumb(D1,"c1")
1

. gen D1 = 3

. mat list D1

D1[2,3]
    c1  c2  c3
r1   1   2   3
r2   4   5   6

. display colnumb(D1,"c1")
type mismatch
r(109);

. display colnumb(matrix(D1),"c1")
1

.

I seem to recall in Stata's copious documentation a discussion of namespaces that explains rules about using or not using the same name for two different things. I cannot find it at the moment, however.