Doubt regarding the dimension of coeff in pca function

Under the description of the function 'pca', it is given that the coeff=pca(X) returns a coeff matrix of p by p matrix if the data matrix X is n by p.

*[coeff = pca(X) returns the principal component coefficients, also known as loadings, for the n-by-p data matrix X. Rows of X correspond to observations and columns correspond to variables. The coefficient matrix is p-by-p.]

However, when I checked with my data I'm getting the coeff dimension as same as data matrix. I also checked by assigning the data matrix using **random numbers. The coeff values will values be p by p ? Could anyone please explain?

========== *Taken from Matlan Tutorial **Obtained by using rand function in Matlab ==========