lm() funtion

Hi,

Currently i am working with the lm() function for some regressions required
for my project.

suppose the formula parameter in that is given by "response ~  terms",after
some testing i found out that when  the number of observations under terms
is less than  the number of columns or features under terms, the
coefficients of regression generated from the linear model object contains
NA which makes it useless.

right now i am in situation where the number of observation is less than
the number of features , can someone help me how to apply regression in
this case.