Neoview Database Administrator's Guide (R2.2)

Output Considerations for Extract Tasks
You can control these aspects of extract output:
• Number of output files or pipes
Use the numberOfOutputFiles (-nf) argument to specify the number of output files or
pipes to be used for the extract task. Increasing the number of output files or pipes can
improve overall performance. Each Reader thread in the Linux Extractor process will process
the data blocks to each output file in round-robin fashion as each block is received from the
Neoview platform.
• Output file or pipe naming
Use the outputItem (-oi) argument to provide the base name of the file or pipe. The
Extractor task appends a stream number to the name. For example, if you specify that the
output file name is /home/data/daily-extract and set the number of output files to
two, the output files will be named /home/data/daily-extract-1 and
/home/data/daily-extract-2.
• Use the concatOutputFiles (-cf) argument to concatenate output files into a single file.
The Extractor task uses the file name you provide and appends it with .out. For example,
if you specify that the output file name is /home/data/daily-extract and specify the
-cf argument, the output file will be named /home/data/daily-extract.out. This
argument is used if you specified the number of output files as more than one. This argument
is not valid for pipes.
When you use this option the Extractor process will still create the output files as described
above. The difference is that when all of the data has been received, an additional step will
be performed in which the output files will be concatenated into a single file. As the data is
copied from an output file it will be deleted.
If the output file already exists, the existing file will be renamed by appending a modification
timestamp to the existing file name, the file is recreated. For example, if the output file prefix
was /home/data/daily-extract and the output files /home/data/daily-extract-1
and /home/data/daily-extract-2 already existed, the current output files would be
renamed to /home/daily-extract-1-nnnnnnnnnnnn and
/home/daily-extract-2-nnnnnnnnnnnn. If an .out file already existed, it would be
renamed in the same way.
Data Formatting
Fixed-Width Output Data
Record format
Following are rules that apply to the format of fixed-width output data:
FormatField
The width of the field will reflect the actual width of the source column, padded with
space characters.
CHAR and VARCHAR
(ISO 8859-1)
The width of the field will reflect four times the number of characters specified for the
source column, padded with blank characters. This is because UTF-8 characters can be
up to 4 bytes in length.
NCHAR, CHAR and
VARCHAR UCS2
6 characters, with trailing blanksSMALLINT
11 characters, with trailing blanksINTEGER
21 characters, with trailing blanksLARGEINT
Output Considerations for Extract Tasks 125