Fixing incorrect sampling_rate

Is there a way for obspy to fix files that have been saved with what is obviously an incorrect sample rate?
For example, here is a report from day-long traces of 4 stations:
LA23: npts=8640000, samp_rate=100.0
LA25: npts=8640001, samp_rate=100.0
LA30: npts=8640000, samp_rate=100.0
LA32: npts=8640000, samp_rate=100.00000762939453

I’m using these traces in eqcorrscan and it throws an error because of that funny sample rate. I’m also wondering if there will be a problem with station LA25 which has one more sample than a day-long file should.
Using trace.resample is not the way to fix this. Can I just read the trace, change trace.stats.npts and proceed? I’d rather fix the input file - we have local copies, are not fetching from remote services.
What is the best way to fix it?

I think for LA32 I would just try

trace.stats.sampling_rate = 100

and for station LA25 you can delete one sample if you need to have the same number of data points for all day files: =[:-1]

Yep, changing stats.sampling_rate is what I decided to do. Since number of samples discrepancy only issues warnings, I’ve left it alone. But if I did change it as you suggest, what about the end time that is stored in stats? I suppose I should recalc it, just in case eqcorrscan cares about it.
But ultimately I’d like to repair the mseed files, because I’m not the only one who wants to use these.

The end time is not stored in the trace and is calculated on demand from start time, npts and sampling rate. Should be no problem.