Error Message r(3900)

[Alexandre Loures]

I'm trying to estimate a logitfe (1,527,680 observations and 378 variables). But it is giving the following error below:


J(): 3900 unable to allocate real <tmp>[1527680,69440]
logit(): - function returned error
<istmt>: - function returned error
r(3900);


The command I'm using is as follows:

logitfe y sur11 sur92 border lang ldist F_*, analytical ieffect(no)

where:

y, sur11, sur92, border lang and F_* are binary variables (taking values 0 and 1)

ldist is the logarithm of the distance between cities

The dummies F_* were created as follows:

xi, pre(F_1) i.ido*i.year
xi, pre(F_2) i.idd*i.year

The settings:
i7 8GB 1T
Windows 10
STATA 14.1

Could anyone help me?

Already thanks!

[Nick Cox]

You're asking, implicitly, for a massive matrix with 106 billion entries, plus some . Stata just can't oblige.

The implication in reverse is that you are trying to estimate some enormously complicated model, so you might want to reconsider that. In fact, Stata is not giving you other options.

[Alexandre Loures]

Thank Nick Cox!