[data-issues] compression error with G.SCZ

Chad chad at iris.washington.edu
Tue Oct 24 10:37:57 PDT 2006

Hi Philip and Matt,

Looks like we have two copies of the G.SCZ data you pointed out.  The older
copy does not include blockette 1000s and includes Steim-1 compression frames
starting at by 48 (not allowed), this data was causing the problems for you.
The newer copy fixes those problems.  Both copies exist in both the POND and
the archive.  So it looks like corrected data was archived but a cleaning of
the older data was not completed.  It was a while ago so we'll have to do some
diagnostics to figure out the extent of this duplication and then formulate a
plan to remove the older stuff.

Thanks for the notification.


Philip Crotwell wrote:
> HI
> One of our SOD users (Matt Fouch) found a problem with G.SCZ, it looks
> like the usual issue of data in the POND being a different, unknown
> compression from data in the archive (steim1). Appears for an event on
> 23 Feb 1995, but I am not sure if it effects events before or after this
> one.
> Let us know what you find.
> thanks,
> Philip
> 2006-10-24 00:25:04,724 WARN 0 Fail.WaveformVector - ECGroup: Event:
> Cyprus Regi
> on | 02/23/1995 21:03:02 GMT | Mag: 5.8 | Depth 15.00 km | (35.039,
> 32.266) G.19820502T00:00:00.000Z.SCZ.  .BHZ.19910927T00:00:00.000Z ,
> G.19820502T00:00:00.000
> Z.SCZ.  .BHN.19910927T00:00:00.000Z , G.19820502T00:00:00.000Z.SCZ. 
> .BHE.199109
> 27T00:00:00.000Z ,  Processor had a system failure
> edu.iris.dmc.seedcodec.SteimException: encoded data length is not
> multiple of 64 bytes (4048)
> Caused by: edu.iris.dmc.seedcodec.SteimException: encoded data length is
> not multiple of 64 bytes (4048)
>         at edu.iris.dmc.seedcodec.Steim1.decode(Steim1.java:53)
>         at edu.iris.dmc.seedcodec.Codec.decompress(Codec.java:121)
>         at
> edu.iris.Fissures.seismogramDC.LocalSeismogramImpl.decodeBits(LocalSeismogramImpl.java:1395)
>         at
> edu.iris.Fissures.seismogramDC.LocalSeismogramImpl.get_as_longs(LocalSeismogramImpl.java:1053)

More information about the data-issues mailing list