"25101.0 and the second 30 mins is 25102.0."
Then missing points will be where
ix = [diff(t) ~= 1 | diff(t) ~= -23] + 1;
Use those indices to infill the missing locations; note will have to work from rear forward or the index locations will change as the vector length grows.
Alternatively, for each day subsection you can find the subsection length as
and any that aren't the multiple of 25 will be short days and then you can interpolate what are found to infill the missing.