NonStop SQL/MP Reference Manual

NonStop SQL/MP Reference Manual—142115
L-26
LOAD Command
{ UNLOADIN | NO UNLOADIN }
(for loading from a magnetic tape only) specifies whether the tape is unloaded when
rewinding occurs. The default is UNLOADIN (the tape is unloaded when rewound).
VARIN
(for loading only from files—not tables—with variable-length, blocked records,
such as those produced using the VAROUT option of COPY) tells LOAD that in-
file contains variable-length, blocked records that begin with a word that contains
the length of the record; the read count equals the value of that length indicator.
You cannot use the TRIM option with VARIN.
key-sequence-option
specifies how to load tables or files with key-sequenced file organization.
{ PARTONLY }
{ PARTOF volume-name }
(for loading files or tables with key-sequenced file organization only) directs SQL to
load only a single partition.
For an SQL table, PARTONLY directs SQL to load the partition specified as out-
file while PARTOF directs SQL to load the partition on volume volume-name.
If you specify PARTOF, you can specify any partition of the table as out-file.
For an Enscribe file, PARTOF directs SQL to load the partition specified as out-
file. volume-name must specify the volume that contains the primary partition.
You cannot use PARTONLY with an Enscribe file.
You cannot use PARTONLY or PARTOF if the table has indexes defined for it. You
cannot use PARTONLY for an audited table if BLOCKSIZE is less than 2KB.
SORTED
(for loading files or tables with key-sequenced file organization only) specifies that
input file records are already sorted in the key-field order of the output file and are
not to be resorted. If you omit this option and the target file is key-sequenced,
LOAD sorts the records before loading the output file.
MAX num-records
(for loading files or tables—but not indexes—with key-sequenced file organization
only) specifies the number of input records as an integer from 0 through
2,147,483,647.
LOAD uses num-records to determine the size of the scratch file to be used by
the SORT process. If you specify the SORTED option, you do not have to specify
the MAX option.
If you underestimate the number of records, the sort can be significantly slower. If
you overestimate the number of records, the cost is small.