Dataloader/MP Reference Manual

Creating a Customized Version of DataLoader/MP
DataLoader/MP Reference Manual—424148-003
5-37
Sorting With DTLSort
For example:
asc 2:4 string, asc 5 for 4, desc 10 for 4 integer
describes the most significant key as a character string in columns 2 through 4 to
be sorted in ascending order; the second key as a character string (the default
type) in columns 5 through 8, again to be sorted in ascending order; and the least
significant key as a 4-byte binary integer in columns 10 through 13 to be sorted in
descending order.
The UPS (User Process Sort) feature of FastSort can speed up sorts of a small
number of records by a factor of 100. If all the records (together with the overhead)
fit into the work space specified, the job gets its speed enhanced by UPS. If the
records do not all fit in the work space, FastSort transparently falls back to the
normal method of sorting.
To request UPS with a 100 kilobyte memory work space for the previous example,
use this string:
asc 2:4 string, asc 5 for 4, desc 10 for 4 integer (UPS 100)
The sort statistics returned by the last call to DTLSort are described by this
structure defined in DTLOADH:
struct DTLSortInfo
{
short MaxRecLen;
long NumBufPages; /* 1KB memory buffer pages */
long NumRecs;
long ElapsedTime; /* In centiseconds */
long ScratchSeeks;
long NumCompares;
long IOWaitTime; /* In centiseconds */
long ScratchFileSize; /* In bytes */
long InitialRuns;
short FirstMergeOrder;
short MergeOrder;
short IntermediatePasses;
long NumDupsRemoved; /* Will be zero */
};
This structure corresponds exactly to the statistics returned by the
SORTMERGESTATISTICS procedure described in the FastSort Manual. Note that
NumDupsRemoved will always be zero because DTLSort directs FastSort not to
remove duplicates.