Outliers, removing 1st and 99th percentiles in dataset

Hi, here is my problem :

I have a dataset called "COMP" which includes many variables such as DY, etc...

The variable "DY" includes many zeros (approx 59300) and some negative numbers while the reste is positive for a total of 90483 observations.

I would like to create a function that removes the rows for the observations below the 1st and above the 99th percentiles.

Here is what i have been trying so far :

COMP = sortrows(COMP,'DY','ascend'); find(COMP.DY == prctile(COMP.DY,1))

The problem is that i dont want to remove all the zeros .. just the 1st percentile ....

Thanks for your help

Answer 1

[Roger Wohlwend]

p = prctile(COMP.DY, [1 99]);
COMP.DY(COMP.DY <= p(1)) = [];
COMP.DY(COMP.DY >= p(2)) = [];

[alex monjambo]

Thanks :) !