How to generate normally distributed time-invariant covariates in panel data?

[Minch Park]

I am trying to generate panel data.
So, I first made the panel and time variables as below.

local N  = 1000 // # units
local T  = 3 // # periods
local NT = `N' * `T'

set obs `NT'
egen pid = seq(), block(`T')
egen tid = seq(), to(`T')

Here, I want to generate normally distributed time-invariant covariates.
Specifically, what I want to generate is

mata:
N = 1000
T = 3
rnormal(N, 2, 0, 1) # J(T, 1, 1)
end

However, I am not sure how to do that in Stata.

(I am trying to do not use loop or reshape or Mata)

[Clyde Schechter]

clear*

local N  = 1000 // # units
local T  = 3 // # periods
local NT = `N' * `T'

set obs `NT'
egen pid = seq(), block(`T')
egen tid = seq(), to(`T')

set seed 1234 // OR WHATEVER YOU LIKE
by pid (tid), sort: gen x = rnormal() if _n == 1
by pid (tid): replace x = x[_n-1] if _n > 1

[Minch Park]

Clyde Schechter It works perfectly! Thank you.