I want to show that the rank and nullity of a matrix $A$ whose entries come from a PID are preserved by when $A$ is multiplied by invertible matrices

i.e If $A=PBQ$, where $P,Q$ are invertible, rank($A$)=rank($B$), nullity($A$)=nullity($B$).

Does anyone know how to show this? I feel like the lack of inverse elements is an issue. Secondly, does this condition hold for other rings in general? i.e commutative rings?

  • $\begingroup$ if P and Q are invertible, they are full rank. $\endgroup$
    – Brad S.
    Nov 5, 2021 at 16:43
  • $\begingroup$ How do you define the rank of a matrix over a PID? $\endgroup$ Nov 6, 2021 at 12:20

1 Answer 1


This is a possible answer:

Let the rank of a matrix be defined to be the maximal number of independent columns of the matrix. This is also equal to the rank of the column space of the matrix.

Refer to: Definition of rank of matrices and nullity of matrices

Suppose $A=PB$ where $P$ is invertible. Suppose columns $\{a_{i_1},...,a_{i_n}\}$ are linearly independent in $A$. Therefore if $$r_1b_{i_1}+...+r_nb_{i_n}=0,$$ then
$$r_1Pb_{i_1}+...+r_nPb_{i_n}=0,$$ which implies that $$r_1a_{i_1}+...+r_na_{i_n}=0.$$ Hence, columns $b_{i_1},..,b_{i_n}$ are linearly independent. Since we can make the same argument for $P^{-1}A=B$, we can deduce that the rank of $A$ is equal to the rank of $B$.

Suppose $A=BQ$ where $Q$ is invertible. This implies that the column space of $A$ is a subset of the column space of $B$ as the columns of $A$ are in the column space of $B$. Since $B=AQ^{-1}$, we can deduce that the $A$ and $B$ have the same column space which implies that they have the same rank.

If $A$ and $B$ have the same dimensions, by the rank-nullity theorem (see link above), the nullity of $A$ equals the nullity of $B$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.