[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[netCDFJava #RUT-379172]: Bad values while reading BUFR files



> Hello,
> Thanks for your answer. I will remember that ucar.bufr.BufrDump is
> deprecated and no longer supported.
> However, I have analyzed bit-to-bit the provided file and it really
> seems that ucar.bufr.BufrDump is right and Netcdf-Java is wrong. For
> instance, let us look at the 'obsRecord.Parcel lifted index (to 500 hPa)
> (see Notes 3 and 4)' field (descriptor in table B : F=0 X=13 Y=42). The
> first bits related to this field in section 4 are :
> 0 1 1 1 0 1 0 0 0 1 0 0 1 1 1 1 1 1 1 1 0 0 0 1 0 0 1 0 ...
> In this sequence :
> Bits 1 to 6 : 0 1 1 1 0 1 : local reference value (29 ie 29-20=9°C)
> Bits 7 to 12 : 0 0 0 1 0 0 : number of bits for each increment (4 bits)
> Bits 13 to 16 : 1 1 1 1 : 1st increment : indicates a missing value
> Bits 17 to 20 : 1 1 1 1 : 2nd increment : indicates a missing value
> Bits 21 to 24 : 0 0 0 1 : 3rd increment : 1 ie 9+1=10°C
> Bits 25 to 28 : 0 0 1 0 : 4th increment : 2 ie 9+2=11°C
> and so on...
> Please could you have a look to this and, if the problem is confirmed,
> give me an approximate delay before it gets resolved ?
> Thanks in advance again.
> Frederic Bacheviller

yes, you're right, i wasnt correctly detecting missing data when the message is 
compressed.

I have a new release (4.0.40) at

  http://www.unidata.ucar.edu/software/netcdf-java/

if you'd like to test and make sure its fixed now.

Thanks for debugging!

John

Ticket Details
===================
Ticket ID: RUT-379172
Department: Support netCDF Java
Priority: Normal
Status: Open