DataLoader/MX Reference Manual (G06.24+)

Specifying File-Related Options for DataLoader/MX
DataLoader/MX Reference Manual—525872-002
4-10
NOREWIND
Use the MAX modifier in conjunction with the INDIRECT modifier when dividing a
single large input file into a number of smaller files of a given size.
NOREWIND
Specify the NOREWIND modifier only for tape files if you want the tape to be left
extended when DataLoader/MX closes it. For an input file, the tape remains at the
point where the last read left it. For an output file, when the close is done,
DataLoader/MX first writes two tape marks, and then repositions the tape between
them. The next program that writes to the tape will create a new file on the tape. If you
want the next program that writes to the tape to extend the file DataLoader/MX was
writing, you must use FUP or some other means to move the tape backwards across
the first tape mark.
If DataLoader/MX terminates abnormally, it ignores the NOREWIND modifier.
NOUNLOAD
Specify the NOUNLOAD modifier only for tape files if you want the tape to be rewound
to the load point and left online when DataLoader/MX closes it.
If DataLoader/MX terminates abnormally, it ignores the NOUNLOAD modifier.
num%
The % modifier specifies the percent of records that are to be processed. num can be
from 0.000 to 100.000, inclusive. For input, only this percentage of the input records
are returned to the application, and the remaining ones are transparently skipped. For
output, only this percentage of the records the application writes are put into the file,
and the remaining ones are transparently discarded. For example:
$ dataload -i="dtldata/nbr1000(text,10%)" -o="outfile(text)"
returns 10 percent of the records in dtldata.nbr1000. The rest are skipped. In this
example, DataLoader/MX returns 10 percent of the records in dtldta/nbr1000, to a
maximum of 20 records:
$ dataload -i="dtldata/nbr1000(text,10%,max=20)" \
-o="outfile(text)"
To prevent any possible synchronization with patterns in the input file, DataLoader/MX
determines whether each record is to be processed by using the built-in random
number generator (DTLRandUnsignedLong). For this reason, the number of records
actually processed might not be exactly the requested percentage, and the difference
can be significant when the number of records is small.
Use this modifier to extract a random subset of a large amount of data to build a
smaller table for testing.