Re: splitting netcdf file into multiple files by date

I believe I found a way to split the file up into
multiple files by data through code since I can't use
a third party tool however the code seems to have an
80 file limit and I'm not sure why.

First I search for a date from a database within a
SearchForTime function

//Get the coordinate variable for the unlimited
dimension
 
ncVarUnlimDim=ncFile->get_var(ncFile->rec_dim()->name());
  
  //Get the type of the unlimited dimension variable
  ncTypeUnlimDim=ncVarUnlimDim->type();
  
  //Get number of values in unlimited dimension
  lNumValsUnlimDim=ncVarUnlimDim->num_vals();
//Scan for date based upon type
  if(ncTypeUnlimDim==ncDouble)
  {
    //double date times may not match exactly created
an epsilon to handle this
    dHalfSec=dEpsilon/(double)timeUnitSec;
   
dRcdTime=_GetTimeAxisValDouble(strDateTime.c_str());
    
    //loop through unlimited dimension to find date
time match or near match
    for(iCount=0; iCount < lNumValsUnlimDim; iCount++)
    {
      if( fabs(dRcdTime -
ncVarUnlimDim->as_double(iCount)) < dHalfSec)
      {
        break;
      }
    }
  }
  else if(ncTypeUnlimDim == ncFloat)

etc........



if there was a match

 //Matching time was found; set time index to this
value
    ncVarUnlimDim->set_cur((long)iCount);
    timeIndex = iCount;

Then the AddTime function is called
Which adds the date to the new file at location 0
//Note all the dimensions, attitbutes, an variables
were already created before this.

Then I have an AddData function which accepts the
original file and the time index found in the
Searchfortime function. However for some reason it
dies after creating 80 files it is dying on the
get_rec
function. Do you all know any limitations on the
get_rec function? Here is the code

for( int i=0 ; i< ncOrgFile->num_vars() ; i++ )
{
  NcVar * ncOrgVar = ncOrgFile->get_var(i);
  NcVar * ncNewVar = ncFile->get_var(i);
  
  // Skip coordinate variables (i.e., variables with
names
        // that match dimension names)
  if( strcmp(ncOrgVar->name(),
ncOrgVar->get_dim(0)->name()) == 0)
  {
    continue;
  }
  NcValues* varOrgVals;
 
  varOrgVals = ncOrgVar->get_rec( orgTimeIndex );
  ncOrgType = ncOrgVar->type();    
  // Insert values at new location
  NcBool rtn;
  if( ncOrgType == ncByte )
  {
    ncbyte *copyVals = new ncbyte[varOrgVals->num()];
    for(int k=0; k < varOrgVals->num() ;k++) 
    { 
      copyVals[k] = varOrgVals->as_ncbyte(k);
    }
    rtn = ncNewVar->put_rec( copyVals, timeIndex);
    delete [] copyVals;
  }
  else if( ncOrgType ==  ncChar )
  {
    etc......
}

--- t.hume@xxxxxxxxxx wrote:
> Hi,
> 
> This is nasty, but it would be quick. Your C++
> program could make a call
> to the ncks utility (which is part of the NetCDF
> operators). In C you
> use the system function ... I guess it is similar or
> the same in C++.
> Compared to using the C++ interface, this will save
> you many many lines
> of code, at the expense of being a very inelegant
> solution.
> 
> Tim Hume
> 
> 
> On Mon, 11 Apr 2005 11:42:20 -0700 (PDT)
> Selina Satterfield <satterfields2000@xxxxxxxxx>
> wrote:
> 
> > I'm new to netCDF development and currently I have
> a
> > large number of netcdf files
> > that contain multiple dimensions, variables , and
> data
> > for a period of time.  Here is a cdl that I
> generated
> > with ncdump of one of the files
> > 
> > netcdf test {
> > dimensions:
> >     lat = 141 ;
> >     lon = 360 ;
> >     time = UNLIMITED ; // (31 currently)
> > variables:
> >     float lat(lat) ;
> >             lat:long_name = "Latitude" ;
> >             lat:units = "degrees_north" ;
> >     float lon(lon) ;
> >             lon:long_name = "Longitude" ;
> >             lon:units = "degrees_east" ;
> >     float time(time) ;
> >             time:long_name = "Time" ;
> >             time:units = "hours since 1997-01-01" ;
> >     float age_rcnt_obs(time, lat, lon) ;
> >             age_rcnt_obs:_FillValue = 999.f ;
> >             age_rcnt_obs:long_name = "Age of Recent
> Observation"
> > ;
> >             age_rcnt_obs:missing_value = 999.f ;
> >             age_rcnt_obs:units = "Hours" ;
> >             age_rcnt_obs:valid_max = 255.f ;
> >             age_rcnt_obs:valid_min = 0.f ;
> >     float anal_temp(time, lat, lon) ;
> >             anal_temp:_FillValue = 9999.f ;
> >             anal_temp:long_name = "Analysis Temperature" ;
> >             anal_temp:missing_value = 999.9f ;
> >             anal_temp:scale_factor = 0.1f ;
> >             anal_temp:units = "Deg. C" ;
> >             anal_temp:valid_max = 70.f ;
> >             anal_temp:valid_min = -10.f ;
> >     float avgrad(time, lat, lon) ;
> >             etc.....;
> > 
> > // global attributes:
> >             :title = "SST 100 KM FIELD" ;
> > data:
> > 
> >  lat = -70, -69, -68, -67, -66, -65, -64, -63,
> -62,
> > -61, -60, -59, -58, -57, 
> >     -56, -55, -54, -53, -52, -51, -50, -49, -48,
> -47,
> > -46, -45, -44, -43, 
> >     etc..............;
> > 
> >  lon = -180, -179, -178, -177, -176, -175, -174,
> -173,
> > -172, -171, -170, 
> >     -169, -168, -167, -166, -165, -164, -163,
> -162,
> > -161, -160, -159, -158, 
> >     -157, -156, -155, -154, -153, -152, -151,
> -150,
> > -149, -148, -147, -146, 
> >     -145, -144, -143, -142, -141, -140, -139,
> -138,
> > -137, -136, -135, -134, 
> >     etc..............;
> > 
> >  time = 60600, 60624, 60648, 60672, 60696, 60720,
> > 60744, 60768, 60792, 60816, 
> >     60840, 60864, 60888, 60912, 60936, 60960,
> 60984,
> > 61008, 61032, 61056, 
> >     61080, 61104, 61128, 61152, 61176, 61200,
> 61224,
> > 61248, 61272, 61296, 
> >     61320 ;
> > 
> >  age_rcnt_obs
> >   .........etc
> > 
> > As you can see with the time there is data for 31
> > days. What I would like to do is create 31
> separate
> > files using the NetCDF C++ Interface. Each file
> would
> > include all dimensions variables attributes and
> only
> > the data for that single day. I can't use a
> utility I
> > need to create this myself since dates and other
> > information needs to be obtained from a database.
> My
> > question is can anyone send a code example of how
> I
> > might do this? Thank you for your time and
> > consideration.
> > 
> > Selina
> > 
> > 
> > 
> > 
> >             
> > __________________________________ 
> > Do you Yahoo!? 
> > Make Yahoo! your home page 
> > http://www.yahoo.com/r/hs
> > 
> 


                
__________________________________ 
Do you Yahoo!? 
Read only the mail you want - Yahoo! Mail SpamGuard. 
http://promotions.yahoo.com/new_mail 

  • 2005 messages navigation, sorted by:
    1. Thread
    2. Subject
    3. Author
    4. Date
    5. ↑ Table Of Contents
  • Search the netcdfgroup archives: