Why fill missing values with 0?

Why the instruction to fill missing values with 0 (the ones remaining after imputing the mean to missing values across all columns)? Since we have to compute correlations, this will bias
the results. Why not dropping the rows with missing values?

hey @sabine.flamand

welcome to DQ community. interesting question.

However, can you please attach the link to the mission where you are tasked with this. It will give us a context of why it was actually asked and we will able to help you better.
Thanks.