XPATFR

Section: User Commands (1)
Updated: November 2000
Index Return to Main Contents

NAME

xpatfr - XPAT Fast-Region index builder

SYNOPSIS

xpatfr [ -v ] [ -m memory k|m ] [ -I index_name ] [ -o output_name ] [ -r region_name | -f Fast_Region_file ] -D data_dictionary

DESCRIPTION

xpatfr builds a Fast-Region index between a Main Index and one or more region indices of the the database specified by data_dictionary. A single region can be specified by region_name. To specify several regions use the Fast_Region_file. The Fast-Region index greatly increases the speed of `within' commands in xpat, where the second argument to the `within' command is region_name (refer to the XPat Reference Manual and Tutorial for further details of the `within' command). The size of the Fast-Region index file can vary widely. This variation is due to the fact that the Fast-Region index is compressed and the amount of compression depends on the nature of the region for which the Fast-Region is built. However, the size of a Fast-Region index file will never exceed 1/32 the size of the Main Index.

The region_name parameter, if used, should be the region name as it is defined in the data_dictionary. Note that if this name contains any spaces it should be surrounded by quotes.

Regions listed in the Fast_Region_file, if used, should be region names as defined in the data_dictionary. These region names may be prefixed with a '*' character. See dbbuild for more information.

OPTIONS

The following options are available:

-v: Specify verbose mode. This option tells xpatfr to print progress messages as it builds the index. By default, xpatfr works quietly, only printing a message if it encounters an error condition .
-m Nk
-m Nm: memory size - use N kilobytes or N megabytes of physical memory for building the Fast-Region index. The minimum amount of memory you must specify is equal to the size of the index for the region named region_name, or for the largest region specified in Fast_Region_file. You can determine this amount by looking at the `Region' section for the region(s) of interest in the data_dictionary. The minimum amount of memory you need to specify, in bytes, is that `Region' section's `Count' field, times 4. Divide this amount by 1024 to specify the memory in KB, or by 1048576 to specify this memory in MB.
: In general, the more memory you give xpatfr, the faster it will run. The maximum amount of memory you can specify is the available physical memory. Because this limitation also applies to xpatbld, the amount of memory that is specified to xpatfr is usually the same as that specied to xpatbld. Refer to the xpatbld(1) man page for a detailed explanation of `available physical memory'.
-I index_name: Specify the index name. By default, xpatfr builds the Fast-Region index for the first Main Index in the data_dictionary. The -I option allows you to specify an alternate Main Index for xpatfr to build the Fast-Region index on (if any others exist).
-o output_name: Specify the file name prefix for the Fast-Region index file. xpatfr produces one file, with the suffix `.fri'. By default xpatfr takes the prefix for that file from the name of the region. However, this default may not be appropriate when the region name has spaces or other non-alphanumeric characters in it (such cases produce filenames that are cumbersome to deal with in Unix). The -o option allows you to specify an alternate filename prefix, such as a short, single word. The output_name is used only if the region_name is specified.

RESTRICTIONS

Both the Main Index and the region index for region_name must both exist before xpatfr is invoked.

EXAMPLES

xpatfr -m 4m -r Headline -D data.dd
xpatfr -v -m 8m -o /text/MyRegion -I first -r "Issue Date" -D text.dd

This document was created by man2html, using the manual pages.
Time: 18:03:38 GMT, March 26, 2001