If input fields are of variable
length (represented by an asterisk, *, in the record section), use
delimiters to separate the fields within the record. Any symbol that
does not appear as part of the data (with the exception of alphabetic and
numeric characters, parentheses, and the asterisk) can function as a delimiter:
carriage returns, commas, and space characters are commonly used.
Ensure that records in the
input data are separated by at least one end-of-line indicator (\n
in UNIX; \r\n in Windows).
If you want to ignore data
from the input text file, describe the field in the record specification,
but do not refer to it in the map specification.
If a database column is not
large enough to receive a source column, the data will be treated as follows:
In the case of character
fields, the entry will be truncated on the right, and a warning message
issued.
In the case of numeric fields,
entries that exceed the acceptable range will produce an error message,
and the offending record will be saved in a file called input-file-name.BAD
for later processing.
SAND CDBMS offers
an extensive scripting language for data loads, allowing for the analysis,
manipulation, and/or conversion of data as it is loaded. Consult the SAND CDBMS Tools Reference Guide
for information about the enhanced NDL++ scripting language.
Loading in Virtual Mode allows the import to be performed
without affecting the real database. Once the import has finished successfully,
the changes can be applied to the database using SHUTDOWN...KEEP CHANGE command.
For more information, consult the Virtual Mode
Operations section of the SAND CDBMS Administration Guide.