Last trace always has no data leading to error in norm determination in linstack = stacking.linstack(traces) about eqcorrscan HOT 9 CLOSED

What have you typed to get the output above?
Is the formatting I have applied correct (see mastering markdown for how to format markdown)?
Is traces a list of streams as required here?

from eqcorrscan.

1. linstack = stacking.linstack(traces) called from multi_trace_plot
2. I do not understand the question
3. Yes list with 16 <obspy.core.stream.Stream objects

from eqcorrscan.

1. I meant the print statements, how did you get this:

1 Trace(s) in Stream:
AF.WHYM.03.SHZ | 2009-03-16T17:32:26.770000Z - 2009-03-16T17:32:26.770000Z | 200.0 Hz, 0 samples
0
matchtr []
/usr/local/lib/python2.7/dist-packages/numpy/core/_methods.py:59: RuntimeWarning: Mean of empty slice.
  warnings.warn("Mean of empty slice.", RuntimeWarning)
/usr/local/lib/python2.7/dist-packages/numpy/core/_methods.py:70: RuntimeWarning: invalid value encountered in double_scalars
  ret = ret.dtype.type(ret / rcount)
norm []
norm_nantonum []
[] 180000

?
2. I edited your comment so that the code output is formatted, does this look correct to you (as in is it what you saw in your terminal), if not, please format it - in general a working (or failing in this case, but reproducible) example is most helpful.
3. Golden, do all traces of all streams have data in them?

from eqcorrscan.

1. I added print commands after every line where the problem occurs:

    for i in range(1, len(streams)):
        for tr in stack:
            tr.plot()
            matchtr = streams[i].select(station=tr.stats.station,
                                        channel=tr.stats.channel)
            print(matchtr)
            print(len(matchtr[0].data))
            if matchtr:
                # Normalize the data before stacking
                if normalize:
	            print('matchtr',matchtr[0].data)

                    norm = matchtr[0].data /\
                        np.sqrt(np.mean(np.square(matchtr[0].data)))
	            print('norm',norm)
                    norm = np.nan_to_num(norm)
	            print('norm_nantonum',norm)
                else:
                    norm = matchtr[0].data
                print(norm,len(tr.data))
                tr.data = np.sum((norm, tr.data), axis=0)
    return stack

3. tr.plot() in 3.rd line shows me the data, so I don't know if it is the streams[i].select part or something else, if I change the order of the streams or remove the last stream, it's the same problem with whatever stream is last in the list, output for second last from same print commands is:

1 Trace(s) in Stream:
AF.WHYM.03.SHZ | 2009-03-15T04:59:58.295000Z - 2009-03-15T05:14:58.290000Z | 200.0 Hz, 180000 samples
180000
matchtr [ 0.00777793  0.05530805  0.18358414 ..., -1.03355332 -1.84595387
  0.00971802]
norm [ 0.00015242  0.00108381  0.00359751 ..., -0.02025346 -0.03617322
  0.00019043]
norm_nantonum [ 0.00015242  0.00108381  0.00359751 ..., -0.02025346 -0.03617322
  0.00019043]
[ 0.00015242  0.00108381  0.00359751 ..., -0.02025346 -0.03617322
  0.00019043] 180000

from eqcorrscan.

It looks like you have a stream with an empty trace in it?
Please try to format code, it makes it much easier to read, it's really easy to do...

from eqcorrscan.

I get the same problem with 2 other completely different datasets, and I already spent many hours on this. It is not as simple as one empty trace, I can assure you that

from eqcorrscan.

Can you email me the streams?

from eqcorrscan.

The streams are not empty but there is something weird about them, I think the first one might be a collection of traces rather than individual traces

from eqcorrscan.

I have now been able to narrow down the problem with the data in traces disappearing: In if realign: in plotting.multi_trace_plot the line _pick.time -= shifts[i] resulted in an error because shifts[i] is a tuple and not a float, so I changed it to _pick.time -= shifts[i][1] which solved the problem but may have caused my problem. If you add the line print("Len Traces data", [len(traces[i].data) for i in range(0,len(traces))]) before and after the if loop it results in the first value for the traces[0].data to be empty (but the stream not being empty because the header info is there) resulting in the problem in linstack. So I may have caused the problem myself through the correction I made, but I don't understand why this hadn't been an issue before.

from eqcorrscan.