This may not improve performance from end-user perspectives since backend requests from MP are parallelized. Removing unnecessary ones I think always makes sense since you keep all the equivalent data that's being queried, but less requests.
In this case, I went through uses of this loader, and made sure we were consistent in the arguments being used. This prevents a query from being issued which results in two separate requests which actually are identical (but can't be combined by data loader due to different arguments).
This may not improve performance from end-user perspectives since backend requests from MP are parallelized. Removing unnecessary ones I think always makes sense since you keep all the equivalent data that's being queried, but less requests.
In this case, I went through uses of this loader, and made sure we were consistent in the arguments being used. This prevents a query from being issued which results in two separate requests which actually are identical (but can't be combined by data loader due to different arguments).