Stata中合法下标调用el()为何非法？矩阵元素调用类型不匹配排查

Answers to Your Stata Matrix Questions #

Question 1: When is el(s,i,j) invalid even if i and j are integers within valid ranges?

The el() function in Stata is built only for numeric matrices. The most common case where it fails with valid indices is:

• If s is a string matrix (like one created with matrix s = J(2,2,"test") or from string datasets), calling el(s,i,j) will throw an error. String matrices require the elstr() function instead to extract elements, since el() can’t handle non-numeric values.

A less common edge case: if you pass a Mata matrix (not a regular Stata matrix) to el(). Mata matrices use their own syntax (e.g., M[i,j]) to access elements, so el() won’t recognize them as valid inputs.

Question 2: Why does di el(P,1,1) throw a "type mismatch" error in my program?

This is a tricky quirk of how Stata parses function calls inside program blocks—even though the exact same command works perfectly in interactive mode! Here’s the breakdown:
When you run noi: di el(P,1,1) within a program define block, Stata’s parser doesn’t properly resolve the matrix reference in the context of the el() function call, leading to a type mismatch.

Luckily, the fixes are simple—use either of these approaches:

1. Use Stata’s native matrix element syntax:

   program define MyProgram
       matrix P = J(1,6,0)
       noi: matrix list P
       noi: di P[1,1]  // This works reliably
   end
   

2. Assign the element to a scalar first:

   program define MyProgram
       matrix P = J(1,6,0)
       noi: matrix list P
       scalar p_val = el(P,1,1)
       noi: di p_val  // This also avoids the error
   end
   

Both methods will correctly output 0 without the type mismatch. The core issue is that el() works seamlessly interactively but needs a small adjustment when used inside program code.

内容的提问来源于stack exchange，提问作者Alexis