Neoview Database Administrator's Guide (R2.2)

Neoview Loader FAQs
1. How do I change the arguments of an existing task? For example, I want to change the
name of the input file.
To update the arguments of an existing task, reconfigure the task (CONFIG_TASK_LOAD)
with the overwrite flag set to Y. (-of Y)
2. How does the update operation work? Can I update only some columns?
The Neoview Loader updates all columns except clustering key columns. The loader uses
the table schema to determine the clustering key columns and uses those columns in the
WHERE clause to update the record.
3. Does the Neoview Loader support updates with deltas? What if the input file contains
values only for the modified columns of a record?
No, the loader does not support deltas. The loader requires all columns to be present in the
input file.
4. What is the transaction size -ts parameter?
Transaction size is the number of records that are sent at one time from the Linux gelserver
to a Neoview gelserver process for insert or update into the target table.
5. What are .bad files?
.bad files contain entries that were rejected. See “Rejected Data Records” (page 116) for a
description of .bad files.
On the reader (Linux) side, a record is written to the .input.bad file if the reader thread
fails to read the record properly. For example, if the record has fewer values than the number
of columns for that table, or there is invalid data for the target column type, it is rejected
and written to the .bad file.
On the writer (Neoview) side, records might be rejected by SQL (for example, due to unique
constraint problems.) The rejected record is passed back to the writer thread on the Linux
side and the record is written to the .sql.bad file.
6. What is the difference between the .bad records created by the reader and writer threads?
If the reader thread rejects the record, the .bad record shows the columns exactly as they
were in the input file. If the writer thread rejects the record, the .bad record shows the
record based on the table schema, so the order of the columns in the .bad record matches
the DDL.
The order of the fields for the rejected records that the writer thread writes to the .bad file
is the same as the order of the columns in the target table, which may be different from the
order of the fields in the original input record. This happens only if you use a format file to
map the input fields to the columns in the target table.
7. Can I reuse a task instance?
Yes, you can use any task instance name, regardless of whether that task instance name was
used for a different task. The task instance name is used in conjunction with the task name
to uniquely identify a running task.
8. How do I specify multiple input files for a single task?
Specify all the data files in the same configure task command. Use the -ii argument for
each data file and specify them in the same gcmd command string. The gelserver will spawn
different reader threads to read from each file.
gcmd -rt CONFIG_TASK_LOAD -ii /home/file1.tbl -ii /home/file2.tbl
………
132 Using the Neoview Loader