NAME

     dslice,  qdslice  -  build  a  new  dataset  consisting   of
     requested  data  slices  from an input dataset or a complete
     transpose of the entire dataset.

SYNOPSIS

     dslice  [  -Nntap  ]  [  -Ootap  ]  [  -ss[start_sample]   |
     -s[start_time]  ]  [  -es[end_sample]  |  -e[end_time]  ]  [
     -is[increment_in_samples] | -i[increment_time] ]  [  -V ]  [
     -? ]

     qdslice  [  -Nntap  ]  [  -Ootap  ]  [  -ss[start_sample]  |
     -s[start_time]  ]  [  -es[end_sample]  |  -e[end_time]  ]  [
     -is[increment_in_samples]   |   -i[increment_time]    ]    [
     -ns[buffer_length_in_samples]]  [ -restart ] [ -R ] [ -V ] [
     -? ]

DESCRIPTION

     The programs dslice and qdslice retrieve a slice (or slices)
     of  data from a USP data set. Dslice requires that the input
     data set reside on disk,  while  qdslice  can  accept  piped
     input  as  long as memory is sufficient to buffer a complete
     input record. While dslice is capable of producing an output
     dataset on a pipeline, the qdslice program requires the out-
     put to reside on disk. The output consists of one record for
     each  requested  slice  and  each  output record has a trace
     count equal to the number of records on the  input  dataset.
     Each trace has a trace length in samples equal to the number
     of traces/record of the input.  ePP A  speed  advantage  for
     qdslice  is  gained by taking advantage of the fact that the
     ouput dataset is on disk. A number of output traces  (speci-
     fied  by  the  -ns parameter) are created in a memory buffer
     before being written to disk. This enables multiple  samples
     to  be read from each input trace before jumping to a sample
     on the next trace. The amount of speed gained in  this  pro-
     cess  varies  significantly by the size of the -ns parameter
     specified.  Since  this  memory  is  allocated  at   program
     startup, a large size may be tried, and, if an error occurs,
     it may be reduced until the program will execute.  If  work-
     ing  on machine where memory is limited, the program fdslice
     may be used to gain similar speed  advantages.  The  primary
     drawback of fdslice is that the output data must be resorted
     (see resorter(1)). The default -ns parameter for qdslice  is
     set  to  the input number of samples per record, so starting
     with this default value and decrementing  it  if  necessary,
     then  abandoning qdslice for fdslice if the needed memory is
     not available seems to be a logical approach. For  extremely
     large datasets, the fact that qdslice requires no intermedi-
     ate disk file must also be weighed into  the  logic  of  the
     user's approach.


     Input consists of specifications  of  the  location  of  the
     first slice, last slice, and intermediate incremental slices
     to take, if any. This may be specified  either  in  time  or
     samples.  If  specified in time, the units are assumed to be
     the same as that used by the sample  interval  in  the  line
     header.  (ie. if the sample interval is in milliseconds, the
     specified times are assumed to be in milliseconds).  Command
     line  arguments  are further detailed in the OPTIONS section
     below.

     If the user wishes to view the slice in  another  direction,
     the program transp, which transposes data, may be used.

OPTIONS

  Command line arguments
     -ss start_sample
          Location  by  sample  number  of  the  first  slice  to
          extract.

     -es end_sample
          Location by sample number of the last slice to extract.

     -is increment_in_samples
          Increment in samples between subsequent slices.

     -s start_time
          Time of the first slice to extract. (  Specified  times
          are relative to the sample interval units. )

     -e end_time
          Time of the last slice to extract.

     -i increment_time
          Time to increment between subsequent slices.

     -ns buffer_length_in_samples  (qdslice only)
          Number of samples from each input trace to  buffer  for
          ouput.      Actual      buffer     size     will     be
          buffer_length_in_samples * number of  traces/record  on
          the input dataset.

     -restart
          The command line argument '-restart' can be used in the
          case  of a qdslice job which did not run to completion.
          The rest of the command line should be the same as that
          used  on the original job. At present, this option does
          not work in conjunction with the  -R  ('reverse'  rota-
          tion) option. Set NOTES below for more details.

     -R   Enter the command  line  argument  '-R'  to  perform  a
          'reverse'  rotation.  (ie.  if you have run one pass of
          qdslice on the data, this will  get  you  back  to  the
          original orientation.)

     -V   Enter the command line argument '-V' to get  additional
          printout.

     -?   Enter the command line  argument  '-?'  to  get  online
          help.   The program terminates after the help screen is
          printed.

NOTES

     If a job using the -R option for 'reverse' rotation does not
     run  to  completion, the output dataset can be scanned using
     sis_verify or scan to detect how much  data  has  been  pro-
     cessed.  A  secondary dataset with the remaining records can
     then be run and gather -S run to re-assemble the two partial
     datasets into a full volume.

AUTHOR(S)

     joe m. wade

SEE ALSO

     fdslice(1),resorter(1),transp(1)

COPYRIGHT

     copyright 2001, Amoco Production Company
               All Rights Reserved
          an affiliate of BP America Inc.