I wanted to know if this should be done or if that is correct approach for data merging.

Here I have a example data from from differential methylation

  chromosome    start   stop    % differential  p-Value q-Value
    chr12   137101  138100  10.0862675  1.94E-06    7.18E-05
    chr6    85749801    85750800    12.77776609 1.35E-08    1.38E-06
    chr6    85749901    85750900    12.77776609 1.35E-08    1.38E-06
    chr6    85750001    85751000    12.77776609 1.35E-08    1.38E-06
    chr6    85750101    85751100    12.77776609 1.35E-08    1.38E-06
    chr6    85750201    85751200    12.77776609 1.35E-08    1.38E-06
    chr6    85750301    85751300    12.77776609 1.35E-08    1.38E-06
    chr6    85750401    85751400    13.61568934 3.39E-08    2.93E-06
    chr8    89092901    89093900    13.5308169  1.76E-15    2.91E-11
    chr8    89093001    89094000    13.15384437 5.68E-15    3.62E-11
    chr8    89093101    89094100    10.46745673 1.25E-13    1.87E-10
    chr8    115870201   115871200   -10.93937088    3.33E-11    1.19E-08

So here let say if we consider chr6 which starts from 2nd row with this 85749801 to 85751300 where if we see the observed % differential they are identical.

In that case if the coordinates are collapsed then would that be correct approach?

Any suggestion or help would be really appreciated

Data processing

methylKit, 
window size: default 1000
step size: script default 1000 
min coverage: default 10
q-value: default 0.01
methylation difference: default 10