{    The purpose of this command is the initial creation of a selection.  Selections
{ are the definitions of what data is to be chosen from the input binary log.
{
{ SELECTION:  This parameter defines the name of the selection.  The name
{       specified here can be used throughout ANABL to refer to the selection.
{       Valid values for this parameter are SCL names.
{
{ STATISTIC_CODE:  The statistic is the primary criterion that defines possible
{       membership in a selection.  Valid values for this parameter are statistic
{       names (ie.  JM3, PM5, etc.) or the keyword NONE.  ANABL does not check
{       to see if the specified name is actually a statistic, but does check to
{       insure the specified value has the correct format for a statistic code.
{       Use of the keyword NONE causes selection of all statistics satisfying
{       the other selection criteria.  That is, the statistic identifier is not
{       a selection criterion.
{
{ TIME:  This optional parameter allows the user to restrict membership in a
{       selection to only those statistics that were emitted on the specified date
{       or within the specified date_time range.  Valid values for this
{       parameter are a single date_time or a range of date_time values.  This
{       parameter is used to specify dates, times, or both dates and times.
{       Values can be specified in any standard SCL date_time string format.
{       The null string (= '') means that date and/or time will not be a
{       selection criterion.  In this case, if all other selection criteria
{       match, a statistic will be selected no matter what date_time it was
{       emitted on.
{
{ CONTINUOUS_DATE_TIME:  The continuous_date_time parameter is used to modify
{       date_time selection.  Time selection is based on the time and
{       continuous_date_time parameters.  The value of the continuous_date_time
{       parameter affects how the values of the time parameter are interpreted
{       in the special case when both date and time are specified as a range.
{       All other cases have unambiguous interpretations, so the value of the
{       continuous_date_time parameter is irrelevant in these cases.  When the
{       value of the continuous_date_time parameter is FALSE, then the
{       specified time range is treated independently from the date range.
{       That is, for every date in the date range, selection selection is based on
{       the time range for that day.  A statistic in the input log satisfies
{       the time selection criteria if it was emitted on a date within the date
{       range and if on that date, it was emitted at a time within the time
{       range.  This is the way one would normally treat time range selection
{       if date selection were based on a single day or not specified at all.
{       When the value of the continuous_date_time parameter is TRUE, then the
{       two ranges, date and time, must be considered together.  The first time
{       in the time range only applies to the first date in the date range.
{       The second time in the time range only applies to the second date in
{       the date range.  A statistic in the input log satisfies the time
{       selection criteria if it occurs after the first time and date
{       combination, but before the second time and date combination.  Thus,
{       while the date the statistic was emitted on must fall withing the date
{       range, the time the statistic was emitted at is only important if the
{       date the statistic was emitted on is an endpoint of the date range.  In
{       fact, the second time range endpoint does not have to be greater than
{       the first.
{
{ DESCRIPTIVE_DATA:  This optional parameter allows the user to restrict
{       membership in a selection to only those statistics having a descriptive
{       data field that matches this specified string.  Values for this
{       parameter must be strings.  The null string (= '') means that the
{       descriptive data field will not be a selection criterion.  In this
{       case, if all other selection criteria match, a statistic will be
{       selected no matter what the value of its descriptive data is, or even
{       whether it does or does not have a descriptive data field.
{
{ JOB_PREDECESSOR:  This optional parameter allows the user to restrict
{       membership in a selection to only those statistics belonging to a job that
{       has previously logged a statistic that belongs to the specified selection.
{       Valid values for this parameter are SCL names.  These names must be
{       previously defined selection names.  A selection can only have one
{       job_predecessor selection.  It may have many successors.  The keyword NONE
{       means that job_predecessor will not be a selection criterion.
{
{ TASK_PREDECESSOR:  This optional parameter allows the user to restrict
{       membership in a selection to only those statistics belonging to a task
{       that has previously logged a statistic that belongs to the specified
{       selection.  Valid values for this parameter are SCL names.  These names must
{       be previously specified selection names.  A selection can only have one
{       task_predecessor selection.  It may have many successors.  The keyword
{       NONE means that task_predecessor will not be a selection criterion.
{
{ SYSTEM_JOB_NAME:  This optional parameter allows the user to restrict
{       membership in a selection to only those statistics that were emitted by
{       the specified job.  Valid value for this parameter is system supplied
{       job name.  The keyword NONE means ANABL will not use the job name as a
{       selection criterion.  In this case, if all other criteria match, a
{       statistic will be selected no matter what the value of its system
{       supplied job name is.
{
{ GLOBAL_TASK_ID:  This optional parameter allows the user to restrict
{       membership in a selection to only statistics that were emitted by the
{       specified task.  Valid value for this parameter is string specifying a
{       global task id.  The null string (='') means ANABL will not use the
{       task name as a selection criterion.  In this case, if all other
{       criteria match, a statistic will be selected no matter what the value
{       of its global task id is.
{
{ STATUS: This parameter specifies the request status.
{       CONDITIONS:
{             pte$date_time_e1
{             pte$invalid_global_task_id
{             pte$redefined_selection
{             pte$task_id_index_out_of_range
{             pte$task_id_seqno_out_of_range
{             pte$undefined_selec_for_predec
